Core Rule Set: gitea's ordinary requests pass by default (closes #198)

The eighth change, internal/waf/gitea.conf, has the fields where gitea
sends a search, a file name or a branch name skip the rules that refused
ordinary gitea requests, on gitea's paths only. Referer skips 944130 too.
A registry's layer reuse skips 932260 on from, the seventh change.

The new setting SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE loads an operator's
own exclusions before the Core Rule Set's rules. A set file that cannot
be read, does not parse or names smallwebwaf's own ids stops the start
in every mode.

A test sends the recorded gitea requests at the defaults, none refused,
and three attacks, all refused.

Model: opus-5-5

README.md

@@ -272,7 +272,7 @@ Every setting the code reads, with its default:
 | `SWWAF_WAF_DISABLED_RULES`                 | `920340,920420,920440,920640,930130,930140`                                    | Core Rule Set rule ids switched off; a list you give replaces the default, so repeat these six to keep them off.                                                    |
 | `SWWAF_WAF_EXEMPT_PATHS`                   | empty                                                                          | Path prefixes the Core Rule Set does not inspect.                                                                                                                   |
 | `SWWAF_WAF_BODY_LIMIT`                     | `off`                                                                          | How much of a request body the Core Rule Set reads, a size up to `64K`; `off` reads none.                                                                           |
-| `SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE`      | unset                                                                          | A file of Core Rule Set exclusions, as Coraza directives, that switch rules off for some fields or paths only; `deploy/gitea/crs-exclusions.conf` is gitea's.       |
+| `SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE`      | unset                                                                          | A file of Core Rule Set exclusions, as Coraza directives, that switch rules off for some fields or paths only.                                                      |
 | `SWWAF_BLOCKLIST_URLS`                     | empty                                                                          | Blocklists to fetch, text files of addresses and netblocks one to a line, such as `https://www.spamhaus.org/drop/drop.txt`.                                         |
 | `SWWAF_BLOCKLIST_REFRESH`                  | `24h`                                                                          | How long after a list was last tried it is fetched again, `1h` or more.                                                                                             |
 | `SWWAF_BLOCKLIST_ACTION`                   | `deny`                                                                         | What a client on a blocklist gets: `deny`, `limit:<percent>` of every limit, or `log`.                                                                              |
@@ -637,14 +637,13 @@ and the status for a size or time limit, such as `413` for a body too large.
   headers of every request, and no body unless `SWWAF_WAF_BODY_LIMIT` is set. In
   front of a code forge, bodies carry what people write and publish, which the
   rules take for attacks.
-- `smallwebwaf` changes the Core Rule Set in seven ways so that it lets ordinary
+- `smallwebwaf` changes the Core Rule Set in eight ways so that it lets ordinary
   requests to a code forge through, such as git over HTTP and sign-in through
-  Git Credential Manager. No setting undoes them; "Attack detection" in
+  Git Credential Manager. On gitea's own paths, its searches, file names and
-  [`SPEC.md`](SPEC.md) lists them.
+  branch names get through too. No setting undoes the changes; "Attack
+  detection" in [`SPEC.md`](SPEC.md) lists them.
 - `SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE` switches single rules off for an app's
-  own fields or paths. gitea's exclusions, in
+  own fields or paths.
-  `deploy/gitea/crs-exclusions.conf`, let its searches, file names and branch
-  names through.
 - Rule files are plain text, one rule to a line, edited by hand and read again
   while `servewaf` runs. The image ships one file of probes no real visitor
   sends (see [`docs/rules.md`](docs/rules.md)).

SPEC.md

@@ -82,7 +82,7 @@ to them.
       of bans is full.
     - `internal/rules`: reads the rule files at start and again when they
       change, and tells which rules a request matches.
-    - `internal/waf`: the Core Rule Set, with the seven changes listed under
+    - `internal/waf`: the Core Rule Set, with the eight changes listed under
       "Attack detection", as Coraza directives.
     - `internal/lookup`: looks clients up through GeoJS or in the lookup
       database.
@@ -587,8 +587,8 @@ sends it.
 
 - `SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE` names a file of exclusions, written as
   Coraza directives. An exclusion switches one rule off for some fields or paths
-  only, where `SWWAF_WAF_DISABLED_RULES` switches it off everywhere.
+  only, where `SWWAF_WAF_DISABLED_RULES` switches it off everywhere. The file
-  `deploy/gitea/crs-exclusions.conf` holds gitea's.
+  holds an app's own exclusions; the eighth change below holds gitea's.
     - `servewaf` loads the file after `smallwebwaf`'s own setup rules and before
       the Core Rule Set's rules. The Core Rule Set puts its own file of
       exclusions that run before the rules in the same place.
@@ -632,7 +632,7 @@ sends it.
     scanner from an ordinary client. In front of gitea, registry and git clients
     get `404` and `401` by design, hundreds of times a minute.
 
-`smallwebwaf` changes the Core Rule Set 4.25.0 in seven ways, since in front of
+`smallwebwaf` changes the Core Rule Set 4.25.0 in eight ways, since in front of
 gitea it would otherwise refuse ordinary requests. The changes hold in front of
 every app, and no setting undoes them:
 
@@ -664,9 +664,10 @@ every app, and no setting undoes them:
     or `docker-build`. In these parameters, what only these three rules would
     catch reaches the app, such as the command `|cat /etc/passwd`. Path
     traversal, SQL and script injection, and PHP, Java and Node.js code are
-    still refused there. Every other parameter keeps all three rules. An app
+    still refused there. Other parameters keep all three rules, except on the
-    that uses one of these parameters as a file on the server, or passes it to a
+    gitea paths the eighth change names. An app that uses one of these
-    shell, must check the value itself. A rule file can refuse it instead.
+    parameters as a file on the server, or passes it to a shell, must check the
+    value itself. A rule file can refuse it instead.
 
 5. Two cookies are left out, and `Referer` keeps fewer rules:
     - The Core Rule Set reads the request without the `gitea_flash` and
@@ -680,19 +681,47 @@ every app, and no setting undoes them:
       only to its own pages. Every other cookie is read in full. An app that
       reads a cookie of either name must check it itself.
     - `Referer` is not checked for a Unix command given without arguments
-      (932340) or for Java starting a process (944110). `Referer` names the page
+      (932340), for Java starting a process (944110) or for a Java class name
-      a request comes from, on everything that page loads and on every link
+      (944130). `Referer` names the page a request comes from, on everything
-      followed from it. The Core Rule Set takes some page addresses for attacks,
+      that page loads and on every link followed from it. The Core Rule Set
-      such as a search for `env`, or OpenJDK's `Runtime.java`, whose path holds
+      takes some page addresses for attacks, such as a search for `env`,
-      both `runtime` and `java.`. All of those requests would be refused.
+      OpenJDK's `Runtime.java`, whose path holds both `runtime` and `java.`, or
-      `Referer` keeps the other rules. An app that passes `Referer` to a shell
+      a file named `java.lang.String.md`. All of those requests would be
-      must check it itself.
+      refused. `Referer` keeps the other rules, except after a gitea search (the
+      eighth change). An app that passes `Referer` to a shell must check it
+      itself.
 6. Responses are not inspected. A raw file from a repository, such as a shell
    script, looks to the response rules like source code leaking from the server.
 7. The `from` parameter of a container registry's upload path,
    `/v2/<image>/blobs/uploads/`, is not checked for command names (932260). When
    a push reuses a layer, `from` names the image the layer comes from. An image
    such as `owner/python3-app` would have that push refused.
+8. On gitea's own paths, the fields where gitea sends a search, a file name or a
+   branch name skip the rules that refused ordinary gitea requests there.
+   `internal/waf/gitea.conf` lists each rule and the request it refused. The
+   fields are:
+    - `q` on gitea's search pages, for code, issues, pull requests, commits,
+      repositories, branches, tags, packages and users. A code search holds
+      code, such as `<script>`, `SELECT * FROM users WHERE` or `../`.
+    - `Referer` of a page opened from such a search, for script and SQL
+      injection and Java code (941110, 941120, 941160, 942100, 944100).
+    - The path of a file's page, raw file, download, blame, history, edit and
+      delete pages, and of the API's call for its contents. Names such as
+      `notes.txt~`, `process.env.d.ts` and `java.lang.String.md` were refused.
+    - A file name in a query parameter: `file` of a commit graph, `path` of the
+      API's list of commits, `body` of the API's call for several files,
+      `template` of a new issue, `redirect_to` of the sign-in page, and `name`
+      of an attachment uploaded through the API.
+    - A branch or tag name in a query parameter, for command names (932260).
+      gitea sends one to delete or restore a branch, to start a release from a
+      tag, to edit a branch protection rule, and to list a directory's last
+      commits.
+    - The `Content-Range` header of an artifact upload from
+      `actions/upload-artifact@v3` (920450).
+
+    Elsewhere these fields keep every rule. The paths are those of gitea at the
+    site root. An app other than gitea that serves one of them must check these
+    fields itself.
 
 The parameter and cookie names in these changes are matched without regard to
 case, as Coraza matches them.

deploy/gitea/compose.yaml

@@ -104,12 +104,11 @@ services:
             # address of an automation fleet on another host here too; without
             # that, it is counted like any client.
             SWWAF_ALLOW_NETS: 10.99.0.128/25,fd10:99:0:0:8000::/65
-            # The Core Rule Set's exclusions for Gitea's searches, file names
+            # In block mode the Core Rule Set refuses everyday Gitea searches
-            # and branch names.
+            # and file names, and 31 refusals in a minute ban the address.
-            SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE: /etc/smallwebwaf/crs-exclusions.conf
+            SWWAF_WAF_MODE: detect
         volumes:
             - servewaf-state:/var/lib/smallwebwaf
-            - ./crs-exclusions.conf:/etc/smallwebwaf/crs-exclusions.conf:ro
         networks:
             - shared
         logging: *logging

deploy/gitea/crs-exclusions.conf

@@ -1,150 +0,0 @@
-# The Core Rule Set's exclusions for Gitea, which servewaf loads from
-# SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE. servewaf loads this file before the Core
-# Rule Set's rules, so it holds only exclusions made while a request is checked:
-# SecAction and SecRule lines with ctl:ruleRemoveTargetById or
-# ctl:ruleRemoveById. Each switches one rule off for the Gitea fields or paths
-# where it refused ordinary Gitea requests. A SecAction applies on every path, a
-# SecRule on the paths it matches. Gitea sends these in the query string:
-#   - q: the text of every search;
-#   - file, path, template, body and redirect_to: a file name, alone or in
-#     JSON or a link;
-#   - name, tag, rule_name and refSubUrl: a branch or tag name.
-# docs/gitea.md says how to add an exclusion.
-
-# Searches.
-# 930110: a search for ../
-SecAction "id:10001,phase:1,pass,nolog,ctl:ruleRemoveTargetById=930110;ARGS:q"
-# 931100: a search for http://127.0.0.1:8080
-SecAction "id:10002,phase:1,pass,nolog,ctl:ruleRemoveTargetById=931100;ARGS:q"
-# 932140: a search for if (a == b) {
-SecAction "id:10003,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932140;ARGS:q"
-# 932235: a search for `make check`
-SecAction "id:10004,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932235;ARGS:q"
-# 932250: a search for ls -la
-SecAction "id:10005,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932250;ARGS:q"
-# 932380: a search for `whoami`
-SecAction "id:10006,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932380;ARGS:q"
-# 933100: a search for <?php echo
-SecAction "id:10007,phase:1,pass,nolog,ctl:ruleRemoveTargetById=933100;ARGS:q"
-# 933130: a search for $_GET[
-SecAction "id:10008,phase:1,pass,nolog,ctl:ruleRemoveTargetById=933130;ARGS:q"
-# 933135: a search for ${VAR}
-SecAction "id:10009,phase:1,pass,nolog,ctl:ruleRemoveTargetById=933135;ARGS:q"
-# 934110: a search for http://127.0.0.1:8080
-SecAction "id:10010,phase:1,pass,nolog,ctl:ruleRemoveTargetById=934110;ARGS:q"
-# 941100: a search for <script>
-SecAction "id:10011,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941100;ARGS:q"
-# 941110: a search for <script>, or a page opened from its results
-SecAction "id:10012,phase:1,pass,nolog,\
-    ctl:ruleRemoveTargetById=941110;ARGS:q,\
-    ctl:ruleRemoveTargetById=941110;REQUEST_HEADERS:Referer"
-# 941120: a page opened from the results of a search for onClick={
-SecAction "id:10013,phase:1,pass,nolog,\
-    ctl:ruleRemoveTargetById=941120;REQUEST_HEADERS:Referer"
-# 941160: a search for <img src=, or a page opened from its results
-SecAction "id:10014,phase:1,pass,nolog,\
-    ctl:ruleRemoveTargetById=941160;ARGS:q,\
-    ctl:ruleRemoveTargetById=941160;REQUEST_HEADERS:Referer"
-# 941390: a search for fetch(
-SecAction "id:10015,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941390;ARGS:q"
-# 942100: a search for SELECT * FROM users WHERE, or a page opened from the
-# results of a search for LIKE '%foo%'
-SecAction "id:10016,phase:1,pass,nolog,\
-    ctl:ruleRemoveTargetById=942100;ARGS:q,\
-    ctl:ruleRemoveTargetById=942100;REQUEST_HEADERS:Referer"
-# 942190: a search for union select
-SecAction "id:10017,phase:1,pass,nolog,ctl:ruleRemoveTargetById=942190;ARGS:q"
-# 942230: a search for if (a == b) {
-SecAction "id:10018,phase:1,pass,nolog,ctl:ruleRemoveTargetById=942230;ARGS:q"
-# 944100: a search for java.lang.Runtime, or a page opened from its results
-SecAction "id:10019,phase:1,pass,nolog,\
-    ctl:ruleRemoveTargetById=944100;ARGS:q,\
-    ctl:ruleRemoveTargetById=944100;REQUEST_HEADERS:Referer"
-
-# Searches, and file, branch and tag names.
-# 930120: a search for package.json, or the commit graph of .gitignore
-SecAction "id:10020,phase:1,pass,nolog,ctl:ruleRemoveTargetById=930120;\
-    ARGS:/^(?:q|file|template|body|refsuburl|name|tag|rule_name)$/"
-# 932160: a search for /bin/sh, or the API's contents of bin/docker-entrypoint
-SecAction "id:10021,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932160;\
-    ARGS:/^(?:q|file|template|body|refsuburl|name|tag|rule_name)$/"
-# 932260: a search for python3, or deleting the branch docker-build
-SecAction "id:10022,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932260;\
-    ARGS:/^(?:q|file|template|body|refsuburl|name|tag|rule_name)$/"
-
-# Searches, and file names in the query string.
-# 932120: a search for Invoke-WebRequest, or the commits of install-script.sh
-SecAction "id:10023,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932120;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 932130: a search for $(pwd), or an issue opened from the template [draft].md
-SecAction "id:10024,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932130;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 934100: a search for process.env, or the commits of process.env.d.ts
-SecAction "id:10025,phase:1,pass,nolog,ctl:ruleRemoveTargetById=934100;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 934130: a search for __proto__, or the commits of __proto__.json
-SecAction "id:10026,phase:1,pass,nolog,ctl:ruleRemoveTargetById=934130;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 941130: the commits of index.xhtml
-SecAction "id:10027,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941130;\
-    ARGS:/^(?:file|path|template|body|redirect_to)$/"
-# 941180: a search for document.cookie, or the commits of document.write.js
-SecAction "id:10028,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941180;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 942140: a search for information_schema, or the commits of
-# information_schema.sql
-SecAction "id:10029,phase:1,pass,nolog,ctl:ruleRemoveTargetById=942140;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 944110: a search for java.lang.Runtime, or the commits of Runtime.java
-SecAction "id:10030,phase:1,pass,nolog,ctl:ruleRemoveTargetById=944110;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/"
-# 944130: a search for java.lang.Runtime, the commits of java.lang.String.md,
-# or any page opened from that file's page, which names it in the Referer
-SecAction "id:10031,phase:1,pass,nolog,\
-    ctl:ruleRemoveTargetById=944130;\
-    ARGS:/^(?:q|file|path|template|body|redirect_to)$/,\
-    ctl:ruleRemoveTargetById=944130;REQUEST_HEADERS:Referer"
-
-# File names in the path of a file's page, raw file, blame, history and edit
-# page, and of the API's calls for its contents.
-# 920500: the page of notes.txt~
-SecRule REQUEST_FILENAME "@rx ^/(?:api/v1/repos/)?[^/]+/[^/]+/(?:src|raw|media|blame|commits|_edit|_delete|contents)/" \
-    "id:10032,phase:1,pass,nolog,ctl:ruleRemoveTargetById=920500;REQUEST_FILENAME"
-# 934100: the page of process.env.d.ts
-SecRule REQUEST_FILENAME "@rx ^/(?:api/v1/repos/)?[^/]+/[^/]+/(?:src|raw|media|blame|commits|_edit|_delete|contents)/" \
-    "id:10033,phase:1,pass,nolog,ctl:ruleRemoveTargetById=934100;REQUEST_FILENAME"
-# 941130: the page of docs/a b.xhtml
-SecRule REQUEST_FILENAME "@rx ^/(?:api/v1/repos/)?[^/]+/[^/]+/(?:src|raw|media|blame|commits|_edit|_delete|contents)/" \
-    "id:10034,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941130;REQUEST_FILENAME"
-# 944130: the page of java.lang.String.md
-SecRule REQUEST_FILENAME "@rx ^/(?:api/v1/repos/)?[^/]+/[^/]+/(?:src|raw|media|blame|commits|_edit|_delete|contents)/" \
-    "id:10035,phase:1,pass,nolog,ctl:ruleRemoveTargetById=944130;REQUEST_FILENAME"
-
-# The file name, in name, of an attachment uploaded through the API to a
-# release, an issue or a comment.
-# 932120: the attachment install-script.sh
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10036,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932120;ARGS:name"
-# 934100: the attachment process.env.d.ts
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10037,phase:1,pass,nolog,ctl:ruleRemoveTargetById=934100;ARGS:name"
-# 934130: the attachment __proto__.json
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10038,phase:1,pass,nolog,ctl:ruleRemoveTargetById=934130;ARGS:name"
-# 941130: the attachment index.xhtml
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10039,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941130;ARGS:name"
-# 941180: the attachment document.write.js
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10040,phase:1,pass,nolog,ctl:ruleRemoveTargetById=941180;ARGS:name"
-# 942140: the attachment information_schema.sql
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10041,phase:1,pass,nolog,ctl:ruleRemoveTargetById=942140;ARGS:name"
-# 944130: the attachment java.lang.String.md
-SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
-    "id:10042,phase:1,pass,nolog,ctl:ruleRemoveTargetById=944130;ARGS:name"
-
-# The header Content-Range of an artifact upload.
-# 920450: an artifact upload from actions/upload-artifact@v3
-SecRule REQUEST_FILENAME "@rx ^/api/actions_pipeline/_apis/pipelines/workflows/[0-9]+/artifacts/[^/]+/upload$" \
-    "id:10043,phase:1,pass,nolog,ctl:ruleRemoveTargetById=920450;REQUEST_HEADERS_NAMES:Content-Range"

docs/gitea.md

@@ -13,8 +13,7 @@ Four containers share one Docker network, the shared network:
   plain HTTP to HTTPS;
 - `rproxy`, in front of gitea;
 - gitea 1.27.3, as its image comes;
-- `servewaf`, with a volume for its state files, and with
+- `servewaf`, with a volume for its state files.
-  `deploy/gitea/crs-exclusions.conf` mounted read-only.
 
 To start it:
 
@@ -31,8 +30,7 @@ To start it:
    `docker run --rm <gitea's image> gitea generate secret SECRET_KEY`, with
    gitea's image as `deploy/gitea/compose.yaml` names it. Do this once per
    install, and keep the key.
-5. Run `docker compose up --detach` in `deploy/gitea`, with
+5. Run `docker compose up --detach` in `deploy/gitea`.
-   `crs-exclusions.conf` beside `compose.yaml`.
 6. Make gitea's first user with `gitea admin user create`, in gitea's container,
    as the user `git`.
 
@@ -48,12 +46,6 @@ under a name of its own, and checks:
   `rproxy`, and that `servewaf` checks them;
 - that gitea logs the visitor's address, not `rproxy`'s;
 - that a request from a container on the shared network skips the checks;
-- that the Core Rule Set refuses none of the ordinary gitea requests in
-  `internal/waf/testdata/gitea-requests.txt`, sent by a signed-in user;
-- that it refuses none of the API calls with JSON bodies;
-- that it refuses no fetch;
-- that it refuses SQL injection, path traversal and script injection in a query
-  string;
 - that with `servewaf` stopped, gitea still answers through `rproxy`.
 
 ## Uploads and downloads
@@ -87,8 +79,11 @@ GiB a day.
 
 ## The checks
 
-The Core Rule Set runs in `block` mode, the default, with gitea's exclusions
+`SWWAF_WAF_MODE` is `detect`. The Core Rule Set then logs a match, with its rule
-(see "The Core Rule Set's exclusions" below).
+ids, and refuses nothing. In `block` mode, the default, it refuses everyday
+gitea searches and file names. The section "What the Core Rule Set refuses in
+`block` mode" below lists them. Each such refusal counts toward the error burst,
+so 31 in one minute ban the address.
 
 Every other check stays at its default. A client may make 120 requests a minute,
 1,500 an hour and 6,000 a day. Archive downloads and blame and history pages are
@@ -158,49 +153,188 @@ within two minutes. Its address is then banned, its running jobs fail, and each
 repeat within a day makes the ban longer. Newer runners ask for work less often.
 Put the runners' addresses in `SWWAF_RATE_LIMIT_EXEMPT_NETS`.
 
-## The Core Rule Set's exclusions
+## What the Core Rule Set refuses in `block` mode
 
-`deploy/gitea/crs-exclusions.conf` switches single Core Rule Set rules off where
+These notes record what the Core Rule Set does in front of gitea in `block`
-they refused ordinary gitea requests. `servewaf` loads it from
+mode, at the other defaults. They are observations, not requirements.
-`SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE`, before the Core Rule Set's rules. Each
+`script/gitea-requests` runs gitea behind `rproxy` and `servewaf` at the
-exclusion covers one rule, in these places only:
+defaults. It sends through them a sample of gitea's ordinary use and the
+requests named below. It reports each request the Core Rule Set refuses, so that
+these notes can be checked again when gitea or the Core Rule Set changes
+version. In `detect` mode, as the setup runs it, the same requests are logged
+with their rule ids, and pass.
 
-- `q`, the text of every search;
+Each refusal answers only that request ("Bans" in `SPEC.md` says what it counts
-- the query parameters that hold a file name: `file`, `path`, `template`, `body`
+toward). The request log names the rule in `waf_rule_ids`, which
-  and `redirect_to`;
+`SWWAF_WAF_DISABLED_RULES` can switch off.
-- the query parameters that hold a branch or tag name: `name`, `tag`,
-  `rule_name` and `refSubUrl`;
-- `Referer`, on a page opened from search results or from a file's page;
-- the path of a file's page, raw file, blame, history and edit page, and of the
-  API's call for its contents;
-- `name`, the file name of an attachment uploaded through the API;
-- the header `Content-Range` of an Actions artifact upload.
 
-gitea reads these as text to search for, as names in a repository, or, for
+### What it lets through
-`Content-Range`, as the part of a file an upload carries. It never runs them as
-SQL, a command or a page's script. So an attack that only the excluded rules
-would catch there reaches gitea and does nothing. Every other field and path
-keeps every rule.
 
-To add an exclusion for a request the Core Rule Set refuses:
+The Core Rule Set lets gitea's ordinary use through, apart from the refusals in
+the sections after this one:
 
-1. Find the request in `servewaf`'s log. Its line gives the path and query, and
+- browsing and views of files in a repository, with their history, blame and the
-   names the rules in `waf_rule_ids`.
+  file tree;
-2. For each rule, add a line with `ctl:ruleRemoveTargetById` for the field it
+- diffs, including their hidden lines and large files, and pull request review;
-   matched. For a field on every path, copy one of the file's `SecAction` lines.
+- git's clone, fetch and push over HTTP;
-   For some paths only, copy one of its `SecRule REQUEST_FILENAME` lines. Give
+- signing in, including the return to the page a visitor came from, and sign-in
-   it a new id. An id from 900000 to 900999 stops the start.
+  with Git Credential Manager, git-credential-oauth or tea;
-3. Above it, write a one-line comment naming the request it refused.
+- the API's calls for a file and its commits;
-4. Add the request to `internal/waf/testdata/gitea-requests.txt`, and run
+- pushing and pulling container images and packages;
-   `make check`.
+- Actions runners, and artifacts uploaded with `actions/upload-artifact@v4`;
-5. Restart `servewaf`.
+- posting issues, pull requests, comments, wiki pages and files saved in the web
+  editor, code included, since no body is read;
+- the page gitea shows after a change, with its message naming what was done.
 
-The Core Rule Set still refuses the page of a file in a directory named `...`,
+### A query string that reads as an attack
-which reads as path traversal.
 
-With `SWWAF_WAF_BODY_LIMIT` set, it also refuses issue text, comments, wiki
+Most often this is a search. A search term is a query parameter, which the Core
-pages and files saved in the web editor that hold shell commands or code.
+Rule Set checks against its lists of system files and commands (see the fourth
+change under "Attack detection" in `SPEC.md`). A release attachment uploaded
+through the API as `docker-compose.yml` is refused the same way. Examples:
 
-`make gitea-requests` runs gitea behind `rproxy` and `servewaf` at the defaults,
+- a name on those lists, such as `package.json`, `.gitignore` or
-without these exclusions. It reports each request the Core Rule Set refuses
+  `docker-compose.yml`;
-there.
+- text that starts with a command name, such as `python3` or `ssh key`, or has
+  one right after a `/` in its first word, such as `feat/docker-support`;
+- a shell command with its options or a system path (`ls -la`, `sed -i`,
+  `/bin/sh`), or a command in backticks;
+- script code (`fetch(`, `${VAR}`, `process.env`), HTML (`<img src=`) or an SQL
+  statement (`SELECT * FROM users WHERE`);
+- a URL naming an IP address or `localhost`.
+
+### A branch, tag or file name that starts with a command name
+
+The Core Rule Set's list of commands (932260) holds names such as `docker-`,
+`python3`, `ansible`, `base64` and `whoami`. A query parameter holding a branch,
+tag or file name that starts with one is refused. So is one that has such a name
+right after a `/`, such as `feat/docker-support`, `fix/docker-build` or
+`renovate/docker-build-push-action-6.x`. gitea's own pages send the name in a
+query parameter that 930120, 932160 and 932260 still check:
+
+- `refSubUrl`, the branch or tag of a directory listing. A listing sends it when
+  it asks for its entries' last commits in a second request. It does so whenever
+  gitea takes more than a second to work them out;
+- `name`, when a branch is deleted or restored on the branches page;
+- `tag`, when a release is started from a tag;
+- `template`, the file of an issue template, when an issue is opened from it;
+- `rule_name`, when a branch protection rule is opened for editing in the
+  repository's settings or has just been saved.
+
+What a visitor sees, for a branch named `docker-build`:
+
+- A directory listing that makes that second request leaves those last commits
+  out and shows an error.
+- The branches page shows an error instead of deleting or restoring the branch.
+- A release started from such a tag gets the `403` answer of `smallwebwaf` in
+  place of its form. So does an issue opened from such a template, and such a
+  rule opened for editing.
+- Saving such a rule stores it. gitea then returns to the branch settings page,
+  which gets the `403` answer, so the save looks as if it failed.
+
+### A file name in a query parameter that trips the code rules
+
+The Core Rule Set's rules for Java, script, PowerShell and SQL find what they
+look for in some file names. This holds too in the ten query parameters that
+930120, 932160 and 932260 do not check (see the fourth change under "Attack
+detection" in `SPEC.md`). Examples:
+
+- `runtime` in OpenJDK's `src/java.base/share/classes/java/lang/Runtime.java`
+  (944110);
+- a Java class, as in `java.lang.String.md` (944130);
+- `document.write.js` (941180);
+- `process.env.d.ts` (934100);
+- `__proto__` (934130);
+- the PowerShell command `Install-Script`, as in `install-script.sh` (932120);
+- a database's name, as in PostgreSQL's `information_schema.sql` (942140).
+
+For such a file, these requests are refused:
+
+- the API's list of its commits (`path`);
+- the API's call for the contents of files that names it in its query (`body`);
+- the commit graph of that file alone (`file`);
+- the hidden lines of its diff (`path`), which stay hidden;
+- the sign-in page, for a visitor who opens a page of the file that needs
+  signing in, such as its edit page (`redirect_to`). The visitor can still sign
+  in by opening the sign-in page itself.
+
+### A path
+
+- A file name ending in `~`, or an `.xhtml` file whose path holds a space.
+- A path that holds what the Core Rule Set takes for a Java class (944130), as
+  `java.lang.String.md` and
+  `META-INF/services/org.apache.commons.logging.LogFactory` do.
+- A path that holds what it takes for Node.js code (934100), as
+  `process.env.d.ts` does.
+- A path with a directory named `...` (930100, 930110), such as `.../notes.txt`.
+
+For such a file, these are refused:
+
+- the page of the file;
+- its raw file;
+- its history and its blame;
+- its edit page.
+
+The directory that holds it lists it, and a clone carries it.
+
+### A container image push that reuses a layer
+
+A push can reuse a layer of an image docker has pushed to that gitea before.
+docker then asks gitea to take the layer from that image, naming it in the query
+parameter `from`, which 932260 checks. The request is refused when the image's
+owner starts with an entry on the list of commands. It is refused too when the
+image's name after the owner starts with one. Examples are `owner/python3-app`,
+`owner/docker-base`, and any image of an owner named `ansible` or `docker-team`.
+
+What happens next depends on docker's image store. A docker that keeps its
+images in containerd's image store fails the push, and fails each new try the
+same way. A new install of Docker Engine 29 does this. A docker that keeps them
+in its own older image store uploads the layer instead. The push then succeeds
+after one refusal for each layer it reuses, and each of those refusals counts
+toward the error burst ("Attack detection" in `SPEC.md`).
+
+### Artifact uploads from `actions/upload-artifact@v3`
+
+This action sends the header `Content-Range`, which the Core Rule Set refuses
+(920450). The action sends two files at a time, does not retry a `403`, and
+sends no further file once one is refused. So an upload is refused once or
+twice, however many files it holds, and the step fails. One upload does not ban
+the runner. The error burst ("Attack detection" in `SPEC.md`) bans it after more
+than 30 such refusals from its address in a minute, as when many jobs upload at
+once. `actions/upload-artifact@v4` does not send the header.
+
+### What body inspection refuses
+
+With `SWWAF_WAF_BODY_LIMIT` set, the Core Rule Set reads request bodies too. In
+front of gitea it then refuses these, until their rule ids, named in the request
+log, are added to `SWWAF_WAF_DISABLED_RULES`:
+
+- issue and comment text, wiki pages and files saved in the web editor that hold
+  shell commands or code (932125, 932235, 932250 and others);
+- package descriptions that show code;
+- PyPI uploads (922130);
+- attachments named like `debug.log` or `config.yml` (932180).
+
+### What `script/gitea-requests` sends
+
+Of the requests above, the script sends:
+
+- for each file named under "A file name in a query parameter that trips the
+  code rules", the API's list of its commits;
+- for `Runtime.java` alone, the commit graph of that file, the hidden lines of
+  its diff, and the API's call for the contents of files that names it;
+- the sign-in page for a visitor who opens the edit page of `document.write.js`;
+- the pages of `java.lang.String.md`,
+  `META-INF/services/org.apache.commons.logging.LogFactory`, `process.env.d.ts`
+  and `.../notes.txt`;
+- one container image push that reuses a layer, which it does not try again.
+
+It does not send the raw file, history, blame or edit page of those four files,
+or the pages of the directories `META-INF/services` and `...`. It also sends a
+short sample of gitea's ordinary use:
+
+- signing in;
+- pages of files and commits, and a diff;
+- the API;
+- a push, and clone and fetch over HTTP;
+- two container image pushes.

internal/waf/gitea.conf

@@ -0,0 +1,236 @@
+# The eighth change: on gitea's own paths, the fields in which gitea sends a
+# search, a file name or a branch name skip the rules that refused ordinary
+# gitea requests there. Each exclusion runs with the request, before the rules
+# it switches off, those of phase 1 included, so this file is loaded before the
+# Core Rule Set's rules. Each comment names a request the rule refused.
+
+# gitea's searches send the text searched for in q: searches of code, of issues
+# and pull requests, of commits, and of repositories, branches, tags, packages
+# and users. q skips these rules there:
+#   930110: ../
+#   930120: .gitignore
+#   931100: http://127.0.0.1:8080
+#   932120: Invoke-WebRequest
+#   932130: $(pwd)
+#   932140: if (a == b) {
+#   932160: /bin/sh
+#   932235: `make check`
+#   932250: ls -la
+#   932260: python3
+#   932380: `whoami`
+#   933100: <?php echo
+#   933130: $_GET[
+#   933135: ${VAR}
+#   934100: process.env
+#   934110: http://127.0.0.1:8080
+#   934130: __proto__
+#   941100: <script>
+#   941110: <script>
+#   941160: <img src=
+#   941180: document.cookie
+#   941390: fetch(
+#   942100: SELECT * FROM users WHERE
+#   942140: information_schema
+#   942190: union select
+#   942230: if (a == b) {
+#   944100: java.lang.Runtime
+#   944110: java.lang.Runtime
+#   944130: java.lang.Runtime
+SecRule REQUEST_FILENAME "@rx ^/(?:explore/(?:repos|code)|issues(?:/search)?|pulls|\
+    user/search_candidates|[^/]+/-/(?:code|packages)|\
+    [^/]+/[^/]+/(?:search|issues(?:/search)?|pulls|branches|tags|\
+    commits/(?:branch|tag|commit)/.+/search)|\
+    api/v1/repos/(?:search|issues/search|[^/]+/[^/]+/issues))$" \
+    "id:900600,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=930110;ARGS:q,\
+    ctl:ruleRemoveTargetById=930120;ARGS:q,\
+    ctl:ruleRemoveTargetById=931100;ARGS:q,\
+    ctl:ruleRemoveTargetById=932120;ARGS:q,\
+    ctl:ruleRemoveTargetById=932130;ARGS:q,\
+    ctl:ruleRemoveTargetById=932140;ARGS:q,\
+    ctl:ruleRemoveTargetById=932160;ARGS:q,\
+    ctl:ruleRemoveTargetById=932235;ARGS:q,\
+    ctl:ruleRemoveTargetById=932250;ARGS:q,\
+    ctl:ruleRemoveTargetById=932260;ARGS:q,\
+    ctl:ruleRemoveTargetById=932380;ARGS:q,\
+    ctl:ruleRemoveTargetById=933100;ARGS:q,\
+    ctl:ruleRemoveTargetById=933130;ARGS:q,\
+    ctl:ruleRemoveTargetById=933135;ARGS:q,\
+    ctl:ruleRemoveTargetById=934100;ARGS:q,\
+    ctl:ruleRemoveTargetById=934110;ARGS:q,\
+    ctl:ruleRemoveTargetById=934130;ARGS:q,\
+    ctl:ruleRemoveTargetById=941100;ARGS:q,\
+    ctl:ruleRemoveTargetById=941110;ARGS:q,\
+    ctl:ruleRemoveTargetById=941160;ARGS:q,\
+    ctl:ruleRemoveTargetById=941180;ARGS:q,\
+    ctl:ruleRemoveTargetById=941390;ARGS:q,\
+    ctl:ruleRemoveTargetById=942100;ARGS:q,\
+    ctl:ruleRemoveTargetById=942140;ARGS:q,\
+    ctl:ruleRemoveTargetById=942190;ARGS:q,\
+    ctl:ruleRemoveTargetById=942230;ARGS:q,\
+    ctl:ruleRemoveTargetById=944100;ARGS:q,\
+    ctl:ruleRemoveTargetById=944110;ARGS:q,\
+    ctl:ruleRemoveTargetById=944130;ARGS:q"
+
+# A page opened from the results of one of those searches has the search's
+# address in Referer, which skips these rules then:
+#   941110: <script>
+#   941120: onClick={
+#   941160: <img src=
+#   942100: LIKE '%foo%'
+#   944100: java.lang.Runtime
+SecRule REQUEST_HEADERS:Referer "@rx ^https?://[^/]+/(?:explore/(?:repos|code)|\
+    issues(?:/search)?|pulls|user/search_candidates|[^/]+/-/(?:code|packages)|\
+    [^/]+/[^/]+/(?:search|issues(?:/search)?|pulls|branches|tags|\
+    commits/(?:branch|tag|commit)/.+/search)|\
+    api/v1/repos/(?:search|issues/search|[^/]+/[^/]+/issues))\?" \
+    "id:900601,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=941110;REQUEST_HEADERS:Referer,\
+    ctl:ruleRemoveTargetById=941120;REQUEST_HEADERS:Referer,\
+    ctl:ruleRemoveTargetById=941160;REQUEST_HEADERS:Referer,\
+    ctl:ruleRemoveTargetById=942100;REQUEST_HEADERS:Referer,\
+    ctl:ruleRemoveTargetById=944100;REQUEST_HEADERS:Referer"
+
+# A file's page, raw file and download, its blame, history, edit and delete
+# pages, and the API's call for its contents name the file in the path, which
+# skips these rules there:
+#   920500: notes.txt~
+#   934100: process.env.d.ts
+#   941130: docs/a b.xhtml
+#   944130: java.lang.String.md
+SecRule REQUEST_FILENAME \
+    "@rx ^/(?:api/v1/repos/)?[^/]+/[^/]+/(?:src|raw|media|blame|commits|_edit|_delete|contents)/" \
+    "id:900610,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=920500;REQUEST_FILENAME,\
+    ctl:ruleRemoveTargetById=934100;REQUEST_FILENAME,\
+    ctl:ruleRemoveTargetById=941130;REQUEST_FILENAME,\
+    ctl:ruleRemoveTargetById=944130;REQUEST_FILENAME"
+
+# The requests below name a file in a query parameter. These rules refused
+# them for the file named:
+#   930120: .gitignore
+#   932120: install-script.sh
+#   932130: [draft].md
+#   932160: bin/docker-entrypoint
+#   932260: docker-compose.yml
+#   934100: process.env.d.ts
+#   934130: __proto__.json
+#   941130: index.xhtml
+#   941180: document.write.js
+#   942140: information_schema.sql
+#   944110: java/lang/Runtime.java
+#   944130: java.lang.String.md
+# Each request skips the rules that refused it. The fourth change already has
+# path and redirect_to skip 930120, 932160 and 932260 on every path.
+
+# A file's commit graph, in file.
+SecRule REQUEST_FILENAME "@rx ^/[^/]+/[^/]+/graph$" \
+    "id:900620,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=930120;ARGS:file,\
+    ctl:ruleRemoveTargetById=932120;ARGS:file,\
+    ctl:ruleRemoveTargetById=932160;ARGS:file,\
+    ctl:ruleRemoveTargetById=932260;ARGS:file,\
+    ctl:ruleRemoveTargetById=934100;ARGS:file,\
+    ctl:ruleRemoveTargetById=934130;ARGS:file,\
+    ctl:ruleRemoveTargetById=941130;ARGS:file,\
+    ctl:ruleRemoveTargetById=941180;ARGS:file,\
+    ctl:ruleRemoveTargetById=942140;ARGS:file,\
+    ctl:ruleRemoveTargetById=944110;ARGS:file,\
+    ctl:ruleRemoveTargetById=944130;ARGS:file"
+
+# The API's list of a file's commits, in path.
+SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/commits$" \
+    "id:900621,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=932120;ARGS:path,\
+    ctl:ruleRemoveTargetById=934100;ARGS:path,\
+    ctl:ruleRemoveTargetById=934130;ARGS:path,\
+    ctl:ruleRemoveTargetById=941130;ARGS:path,\
+    ctl:ruleRemoveTargetById=941180;ARGS:path,\
+    ctl:ruleRemoveTargetById=942140;ARGS:path,\
+    ctl:ruleRemoveTargetById=944110;ARGS:path,\
+    ctl:ruleRemoveTargetById=944130;ARGS:path"
+
+# The API's call for the contents of several files, in body, as JSON.
+SecRule REQUEST_FILENAME "@rx ^/api/v1/repos/[^/]+/[^/]+/file-contents$" \
+    "id:900622,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=930120;ARGS:body,\
+    ctl:ruleRemoveTargetById=932120;ARGS:body,\
+    ctl:ruleRemoveTargetById=932160;ARGS:body,\
+    ctl:ruleRemoveTargetById=934100;ARGS:body,\
+    ctl:ruleRemoveTargetById=934130;ARGS:body,\
+    ctl:ruleRemoveTargetById=941130;ARGS:body,\
+    ctl:ruleRemoveTargetById=941180;ARGS:body,\
+    ctl:ruleRemoveTargetById=942140;ARGS:body,\
+    ctl:ruleRemoveTargetById=944110;ARGS:body,\
+    ctl:ruleRemoveTargetById=944130;ARGS:body"
+
+# An issue opened from a template, in template.
+SecRule REQUEST_FILENAME "@rx ^/[^/]+/[^/]+/issues/new$" \
+    "id:900623,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=930120;ARGS:template,\
+    ctl:ruleRemoveTargetById=932120;ARGS:template,\
+    ctl:ruleRemoveTargetById=932130;ARGS:template,\
+    ctl:ruleRemoveTargetById=932160;ARGS:template,\
+    ctl:ruleRemoveTargetById=932260;ARGS:template,\
+    ctl:ruleRemoveTargetById=934100;ARGS:template,\
+    ctl:ruleRemoveTargetById=934130;ARGS:template,\
+    ctl:ruleRemoveTargetById=941130;ARGS:template,\
+    ctl:ruleRemoveTargetById=941180;ARGS:template,\
+    ctl:ruleRemoveTargetById=942140;ARGS:template,\
+    ctl:ruleRemoveTargetById=944110;ARGS:template,\
+    ctl:ruleRemoveTargetById=944130;ARGS:template"
+
+# The sign-in page on the way to a file's edit page, in redirect_to.
+SecRule REQUEST_FILENAME "@rx ^/user/login$" \
+    "id:900624,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=932120;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=932130;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=934100;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=934130;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=941130;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=941180;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=942140;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=944110;ARGS:redirect_to,\
+    ctl:ruleRemoveTargetById=944130;ARGS:redirect_to"
+
+# An attachment uploaded through the API to a release, an issue or a comment,
+# in name.
+SecRule REQUEST_FILENAME \
+    "@rx ^/api/v1/repos/[^/]+/[^/]+/(?:releases|issues|issues/comments)/[0-9]+/assets$" \
+    "id:900625,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=930120;ARGS:name,\
+    ctl:ruleRemoveTargetById=932120;ARGS:name,\
+    ctl:ruleRemoveTargetById=932260;ARGS:name,\
+    ctl:ruleRemoveTargetById=934100;ARGS:name,\
+    ctl:ruleRemoveTargetById=934130;ARGS:name,\
+    ctl:ruleRemoveTargetById=941130;ARGS:name,\
+    ctl:ruleRemoveTargetById=941180;ARGS:name,\
+    ctl:ruleRemoveTargetById=942140;ARGS:name,\
+    ctl:ruleRemoveTargetById=944130;ARGS:name"
+
+# The requests below name a branch or tag in a query parameter, which skips
+# the list of command names (932260) there. It refused names such as
+# docker-build, python3-migration and feat/docker-support.
+
+# Deleting or restoring a branch, in name.
+SecRule REQUEST_FILENAME "@rx ^/[^/]+/[^/]+/branches/(?:delete|restore)$" \
+    "id:900630,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932260;ARGS:name"
+
+# A release started from a tag, in tag.
+SecRule REQUEST_FILENAME "@rx ^/[^/]+/[^/]+/releases/new$" \
+    "id:900631,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932260;ARGS:tag"
+
+# A branch protection rule opened for editing, in rule_name.
+SecRule REQUEST_FILENAME "@rx ^/[^/]+/[^/]+/settings/branches/edit$" \
+    "id:900632,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932260;ARGS:rule_name"
+
+# A directory listing's last commits, in refSubUrl.
+SecRule REQUEST_FILENAME "@rx ^/[^/]+/[^/]+/lastcommit/" \
+    "id:900633,phase:1,pass,nolog,ctl:ruleRemoveTargetById=932260;ARGS:refSubUrl"
+
+# An artifact upload from actions/upload-artifact@v3 sends the header
+# Content-Range, which the Core Rule Set refused (920450).
+SecRule REQUEST_FILENAME \
+    "@rx ^/api/actions_pipeline/_apis/pipelines/workflows/[0-9]+/artifacts/[^/]+/upload$" \
+    "id:900640,phase:1,pass,nolog,\
+    ctl:ruleRemoveTargetById=920450;REQUEST_HEADERS_NAMES:Content-Range"

internal/waf/testdata/gitea-requests.txt

@@ -1,8 +1,7 @@
-# Ordinary gitea requests that the Core Rule Set must let through with
+# Ordinary gitea 1.27 requests that the Core Rule Set must let through at the
-# deploy/gitea/crs-exclusions.conf. Each line is a path and query, and may end
+# defaults. Each line is a path and query, and may end with one header after a
-# with one header after a space. The tests of internal/waf inspect each as a GET,
+# space. The tests of internal/waf inspect each as a GET. The Core Rule Set reads
-# and script/gitea-setup sends each to gitea through the setup as a GET. The Core
+# the query and the headers the same way for any method.
-# Rule Set reads the query and the headers the same way for any method.
 
 # Searches with ordinary words and symbols.
 /owner/repo/search?q=package.json
@@ -51,24 +50,30 @@
 /explore/repos?q=%3Cimg%20src%3D
 /explore/code?q=SELECT%20%2A%20FROM%20users%20WHERE
 /explore/code?q=%3Cimg%20src%3D
+/owner/-/code?q=SELECT%20%2A%20FROM%20users%20WHERE
+/owner/-/code?q=%3Cimg%20src%3D
 /owner/repo/issues?type=all&state=open&q=SELECT%20%2A%20FROM%20users%20WHERE
 /owner/repo/issues?type=all&state=open&q=%3Cimg%20src%3D
+/owner/repo/issues/search?q=SELECT%20%2A%20FROM%20users%20WHERE
+/owner/repo/issues/search?q=%3Cimg%20src%3D
 /owner/repo/pulls?q=SELECT%20%2A%20FROM%20users%20WHERE
 /owner/repo/pulls?q=%3Cimg%20src%3D
 /issues?type=your_repositories&q=SELECT%20%2A%20FROM%20users%20WHERE
 /issues?type=your_repositories&q=%3Cimg%20src%3D
+/issues/search?q=SELECT%20%2A%20FROM%20users%20WHERE
+/issues/search?q=%3Cimg%20src%3D
+/pulls?type=your_repositories&q=SELECT%20%2A%20FROM%20users%20WHERE
+/pulls?type=your_repositories&q=%3Cimg%20src%3D
 /owner/repo/commits/branch/main/search?q=SELECT%20%2A%20FROM%20users%20WHERE
 /owner/repo/commits/branch/main/search?q=%3Cimg%20src%3D
-/owner/repo/wiki/search?q=SELECT%20%2A%20FROM%20users%20WHERE
-/owner/repo/wiki/search?q=%3Cimg%20src%3D
 /owner/repo/branches?q=SELECT%20%2A%20FROM%20users%20WHERE
 /owner/repo/branches?q=%3Cimg%20src%3D
-/owner/repo/releases?q=SELECT%20%2A%20FROM%20users%20WHERE
+/owner/repo/tags?q=SELECT%20%2A%20FROM%20users%20WHERE
-/owner/repo/releases?q=%3Cimg%20src%3D
+/owner/repo/tags?q=%3Cimg%20src%3D
 /owner/-/packages?q=SELECT%20%2A%20FROM%20users%20WHERE
 /owner/-/packages?q=%3Cimg%20src%3D
-/user/search?q=SELECT%20%2A%20FROM%20users%20WHERE
+/user/search_candidates?q=SELECT%20%2A%20FROM%20users%20WHERE
-/user/search?q=%3Cimg%20src%3D
+/user/search_candidates?q=%3Cimg%20src%3D
 /api/v1/repos/search?q=SELECT%20%2A%20FROM%20users%20WHERE
 /api/v1/repos/search?q=%3Cimg%20src%3D
 /api/v1/repos/issues/search?q=SELECT%20%2A%20FROM%20users%20WHERE
@@ -155,6 +160,25 @@
 /owner/repo/_edit/main/notes.txt~
 /api/v1/repos/owner/repo/contents/notes.txt~
 
+# The download, blame, history and delete pages of the files above whose names
+# the Core Rule Set checks in the path.
+/owner/repo/media/branch/main/notes.txt~
+/owner/repo/blame/branch/main/notes.txt~
+/owner/repo/commits/branch/main/notes.txt~
+/owner/repo/_delete/main/notes.txt~
+/owner/repo/media/branch/main/process.env.d.ts
+/owner/repo/blame/branch/main/process.env.d.ts
+/owner/repo/commits/branch/main/process.env.d.ts
+/owner/repo/_delete/main/process.env.d.ts
+/owner/repo/media/branch/main/docs/a%20b.xhtml
+/owner/repo/blame/branch/main/docs/a%20b.xhtml
+/owner/repo/commits/branch/main/docs/a%20b.xhtml
+/owner/repo/_delete/main/docs/a%20b.xhtml
+/owner/repo/media/branch/main/java.lang.String.md
+/owner/repo/blame/branch/main/java.lang.String.md
+/owner/repo/commits/branch/main/java.lang.String.md
+/owner/repo/_delete/main/java.lang.String.md
+
 # File names in the query string: the commit graph of a file, the API's lists
 # of its commits and of its contents, an issue opened from a template, and the
 # sign-in page on the way to its edit page.
@@ -281,35 +305,41 @@
 # An artifact upload from actions/upload-artifact@v3, which sends Content-Range.
 /api/actions_pipeline/_apis/pipelines/workflows/1/artifacts/9b7c1f3a/upload?itemPath=artifact%2Freport.txt Content-Range: bytes 0-9/10
 
-# Branch and tag names in the query string: deleting a branch, starting a
+# Branch and tag names in the query string: deleting and restoring a branch,
-# release from a tag, a branch protection rule, a directory listing's last
+# starting a release from a tag, a branch protection rule, a directory listing's
-# commits, and a container image push that reuses a layer.
+# last commits, and a container image push that reuses a layer.
-/owner/repo/branches?name=docker-build
+/owner/repo/branches/delete?name=docker-build&page=1
+/owner/repo/branches/restore?branch_id=1&name=docker-build&page=1
 /owner/repo/releases/new?tag=docker-build
 /owner/repo/settings/branches/edit?rule_name=docker-build
 /owner/repo/lastcommit/main/src?refSubUrl=branch%2Fdocker-build
 /v2/owner/app/blobs/uploads/?mount=sha256%3A0000&from=owner%2Fdocker-build
-/owner/repo/branches?name=feat%2Fdocker-support
+/owner/repo/branches/delete?name=feat%2Fdocker-support&page=1
+/owner/repo/branches/restore?branch_id=1&name=feat%2Fdocker-support&page=1
 /owner/repo/releases/new?tag=feat%2Fdocker-support
 /owner/repo/settings/branches/edit?rule_name=feat%2Fdocker-support
 /owner/repo/lastcommit/main/src?refSubUrl=branch%2Ffeat%2Fdocker-support
 /v2/owner/app/blobs/uploads/?mount=sha256%3A0000&from=owner%2Ffeat%2Fdocker-support
-/owner/repo/branches?name=renovate%2Fdocker-build-push-action-6.x
+/owner/repo/branches/delete?name=renovate%2Fdocker-build-push-action-6.x&page=1
+/owner/repo/branches/restore?branch_id=1&name=renovate%2Fdocker-build-push-action-6.x&page=1
 /owner/repo/releases/new?tag=renovate%2Fdocker-build-push-action-6.x
 /owner/repo/settings/branches/edit?rule_name=renovate%2Fdocker-build-push-action-6.x
 /owner/repo/lastcommit/main/src?refSubUrl=branch%2Frenovate%2Fdocker-build-push-action-6.x
 /v2/owner/app/blobs/uploads/?mount=sha256%3A0000&from=owner%2Frenovate%2Fdocker-build-push-action-6.x
-/owner/repo/branches?name=python3-migration
+/owner/repo/branches/delete?name=python3-migration&page=1
+/owner/repo/branches/restore?branch_id=1&name=python3-migration&page=1
 /owner/repo/releases/new?tag=python3-migration
 /owner/repo/settings/branches/edit?rule_name=python3-migration
 /owner/repo/lastcommit/main/src?refSubUrl=branch%2Fpython3-migration
 /v2/owner/app/blobs/uploads/?mount=sha256%3A0000&from=owner%2Fpython3-migration
-/owner/repo/branches?name=ansible-roles
+/owner/repo/branches/delete?name=ansible-roles&page=1
+/owner/repo/branches/restore?branch_id=1&name=ansible-roles&page=1
 /owner/repo/releases/new?tag=ansible-roles
 /owner/repo/settings/branches/edit?rule_name=ansible-roles
 /owner/repo/lastcommit/main/src?refSubUrl=branch%2Fansible-roles
 /v2/owner/app/blobs/uploads/?mount=sha256%3A0000&from=owner%2Fansible-roles
-/owner/repo/branches?name=whoami
+/owner/repo/branches/delete?name=whoami&page=1
+/owner/repo/branches/restore?branch_id=1&name=whoami&page=1
 /owner/repo/releases/new?tag=whoami
 /owner/repo/settings/branches/edit?rule_name=whoami
 /owner/repo/lastcommit/main/src?refSubUrl=branch%2Fwhoami

internal/waf/waf.go

@@ -1,6 +1,6 @@
 // Package waf runs the OWASP Core Rule Set 4.25.0, through Coraza, on the
 // method, the URL with its query and the headers of a request, and on its
-// body while SWWAF_WAF_BODY_LIMIT is set, with the seven changes smallwebwaf
+// body while SWWAF_WAF_BODY_LIMIT is set, with the eight changes smallwebwaf
 // makes to it and the exclusions in SWWAF_CORE_RULE_SET_EXCLUSIONS_FILE, as
 // "Attack detection" under "Settings" in SPEC.md describes them. It reads no
 // response.
@@ -12,6 +12,7 @@
 package waf
 
 import (
+	_ "embed"
 	"errors"
 	"fmt"
 	"io"
@@ -31,7 +32,8 @@ import (
 
 // directives are the Core Rule Set as smallwebwaf runs it, with the
 // paranoia level for %d, bodyDirectives for the first %s while
-// SWWAF_WAF_BODY_LIMIT is set, and the exclusions file for the second.
+// SWWAF_WAF_BODY_LIMIT is set, the exclusions file for the second, and
+// giteaExclusions for the third.
 // Each rule smallwebwaf adds has an id from 900000 to 900999, the ids the
 // Core Rule Set keeps for the rules that set it up, which
 // SWWAF_WAF_DISABLED_RULES and the exclusions file refuse, so that no
@@ -79,6 +81,9 @@ SecRule ARGUMENTS_LIMIT_REACHED "@eq 1" "id:900300,phase:2,pass,\
 # before the rules of phase 1 too.
 %s
 
+# The eighth change comes before the rules for the same reason.
+%s
+
 # The sixth: only the rules for requests are loaded, and no response is
 # inspected.
 Include @owasp_crs/REQUEST-*.conf
@@ -106,10 +111,12 @@ SecRuleUpdateTargetById 932260 "!ARGS:path|!ARGS:files|!ARGS:skip-to|\
     !ARGS:artifactName|!ARGS:redirect_to"
 
 # The fifth, for Referer: it is not checked for a Unix command without
-# arguments, or for Java starting a process. The cookies are left out in
+# arguments, for Java starting a process, or for a Java class name. 944130
-# Inspect.
+# refused every page opened from the page of java.lang.String.md. The
+# cookies are left out in Inspect.
 SecRuleUpdateTargetById 932340 "!REQUEST_HEADERS:Referer"
 SecRuleUpdateTargetById 944110 "!REQUEST_HEADERS:Referer"
+SecRuleUpdateTargetById 944130 "!REQUEST_HEADERS:Referer"
 
 # The seventh: a container image push that reuses a layer names the image
 # it takes the layer from in the query parameter from, which is not
@@ -161,6 +168,13 @@ SecRule MULTIPART_STRICT_ERROR "!@eq 0" "id:900450,phase:2,pass,\
     severity:'CRITICAL',setvar:'tx.inbound_anomaly_score_pl1=+5'"
 `
 
+// giteaExclusions are the eighth change: the fields of gitea's own
+// requests that skip the rules that refused them there, as Coraza
+// directives.
+//
+//go:embed gitea.conf
+var giteaExclusions string
+
 // cookiesNotRead are the cookies the Core Rule Set reads a request
 // without, the rest of the fifth change.
 //
@@ -231,7 +245,7 @@ type CoreRuleSet struct {
 	bodyLimit int64
 }
 
-// New returns the Core Rule Set with the seven changes, at params'
+// New returns the Core Rule Set with the eight changes, at params'
 // paranoia level, without the rules it switches off, reading request
 // bodies up to params' limit, and with params' exclusions file loaded
 // before the Core Rule Set's rules. An exclusions file that cannot be
@@ -259,7 +273,7 @@ func New(params Params) (*CoreRuleSet, error) {
 		}
 	}
 
-	text := fmt.Sprintf(directives, params.ParanoiaLevel, body, exclusions)
+	text := fmt.Sprintf(directives, params.ParanoiaLevel, body, exclusions, giteaExclusions)
 
 	if len(params.DisabledRules) > 0 {
 		ids := make([]string, len(params.DisabledRules))

internal/waf/waf_test.go

@@ -120,7 +120,7 @@ func atDefaults(t *testing.T) *waf.CoreRuleSet {
 }
 
 // wantChange checks that crs lets through passes, a gitea request one of
-// the seven changes is for, and still finds result in refused, a request
+// the eight changes is for, and still finds result in refused, a request
 // like it that the change is not for.
 func wantChange(
 	t *testing.T, crs *waf.CoreRuleSet, passes, refused request, result waf.Result,
@@ -309,13 +309,14 @@ func TestCookiesGiteaFlashAndRedirectToAreNotRead(t *testing.T) {
 		matched(930120))
 }
 
-func TestRefererIsNotCheckedForACommandOrJavaStartingAProcess(t *testing.T) {
+func TestRefererIsNotCheckedForACommandOrJava(t *testing.T) {
 	t.Parallel()
 
 	const (
 		search      = "https://git.example/explore/repos?q=env"
 		runtimeJava = "https://git.example/openjdk/jdk/src/branch/master/src/" +
 			"java.base/share/classes/java/lang/Runtime.java"
+		stringMD = "https://git.example/owner/repo/src/branch/main/java.lang.String.md"
 	)
 
 	crs := atDefaults(t)
@@ -324,6 +325,8 @@ func TestRefererIsNotCheckedForACommandOrJavaStartingAProcess(t *testing.T) {
 		matched(932340))
 	wantChange(t, crs, get("/", "Referer: "+runtimeJava),
 		get("/", "X-Page: "+runtimeJava), matched(944110))
+	wantChange(t, crs, get("/", "Referer: "+stringMD), get("/", "X-Page: "+stringMD),
+		matched(944130))
 
 	// It is still checked for script and SQL injection.
 	wantResult(t, crs,
@@ -464,9 +467,9 @@ func TestExclusionsFileThatCannotBeUsedIsAnError(t *testing.T) {
 	}
 }
 
-// giteaRequests returns the recorded ordinary gitea requests in testdata,
+// giteaRequests returns the recorded ordinary gitea requests in testdata.
-// which script/gitea-setup sends too. Each has the Referer of a page that
+// Each has the Referer of a page that names the same, as a page opened
-// names the same, as a page opened from it has.
+// from it has.
 func giteaRequests(t *testing.T) []request {
 	t.Helper()
 
@@ -496,30 +499,30 @@ func giteaRequests(t *testing.T) []request {
 	return requests
 }
 
-func TestGiteaExclusionsLetOrdinaryGiteaRequestsThroughAndRefuseAttacks(t *testing.T) {
+func TestOrdinaryGiteaRequestsPassAndAttacksAreRefused(t *testing.T) {
 	t.Parallel()
 
-	crs, err := withExclusions(filepath.Join("..", "..", "deploy", "gitea",
-		"crs-exclusions.conf"))
-	if err != nil {
-		t.Fatalf("load gitea's exclusions: %v", err)
-	}
-
 	const threshold = 5 // SWWAF_WAF_ANOMALY_THRESHOLD's default
 
+	crs := atDefaults(t)
+
 	for _, r := range giteaRequests(t) {
 		if got := inspect(t, crs, r); got.Score >= threshold {
 			t.Errorf("%s %q refused: %+v", r.target, r.headers[1:], got)
 		}
 	}
 
+	// In q too, off gitea's search pages.
+	for _, parameter := range []string{"id", "q"} {
 		for _, attack := range []string{
-		"/owner/repo?id=1'%20OR%20'1'='1",
+			"1'%20OR%20'1'='1",
-		"/owner/repo?id=..%2Fcustom%2Fconf%2Fapp.ini",
+			"..%2Fcustom%2Fconf%2Fapp.ini",
-		"/owner/repo?id=%3Cscript%3Ealert(1)%3C%2Fscript%3E",
+			"%3Cscript%3Ealert(1)%3C%2Fscript%3E",
 		} {
-		if got := inspect(t, crs, get(attack)); got.Score < threshold {
+			target := "/owner/repo?" + parameter + "=" + attack
-			t.Errorf("%s let through: %+v", attack, got)
+			if got := inspect(t, crs, get(target)); got.Score < threshold {
+				t.Errorf("%s let through: %+v", target, got)
+			}
 		}
 	}
 }

script/gitea-requests

@@ -4,8 +4,7 @@
 # default but these: SWRP_UPSTREAM_URL names gitea's container, SWRP_WAF_BACKEND
 # names servewaf's, and SWWAF_LOG_REQUEST_HEADERS adds the header that labels a
 # request. Then send through rproxy the gitea requests that reading gitea's
-# source found the Core Rule Set refuses without the exclusions in
+# source found the Core Rule Set refuses, which docs/gitea.md records, and a
-# deploy/gitea/crs-exclusions.conf, and a
 # short sample of gitea's ordinary use: signing in, pages of files and commits,
 # a diff, the API, a push, clone and fetch over HTTP, and two container image
 # pushes. Last, print each request servewaf logged, with what it did with it and

script/gitea-setup

@@ -5,12 +5,6 @@
 #     rproxy, and that servewaf checks them;
 #   - that gitea logs the visitor's address, not rproxy's;
 #   - that a request from a container on the shared network skips the checks;
-#   - that the Core Rule Set refuses none of the ordinary gitea requests in
-#     internal/waf/testdata/gitea-requests.txt;
-#   - that it refuses none of the API calls with JSON bodies;
-#   - that it refuses no fetch;
-#   - that it refuses SQL injection, path traversal and script injection in a
-#     query string;
 #   - that with servewaf stopped, gitea still answers through rproxy.
 # It publishes no port of traefik's, since this host's ports 80 and 443 may be
 # taken. It reaches traefik at its address on the shared network instead, so
@@ -142,71 +136,6 @@ passed_through() {
         "\"client_ip\":\"$gateway\"" '"action":"forward"' '"counts":' '"waf_score":'
 }
 
-# refused: print each request servewaf logged as refused, for any reason.
-refused() {
-    compose logs --no-color --no-log-prefix servewaf 2>&1 |
-        grep -F '"servewaf_line":"report"' | grep -v -F '"action":"forward"' || true
-}
-
-# ordinary_requests: sign in as the owner, save a file and open an issue
-# through the API with JSON bodies, then send, signed in, each request in
-# internal/waf/testdata/gitea-requests.txt, with the header its line ends with,
-# if any, and that request's own address as its Referer, as a page opened from it
-# has. One a second keeps them under the default limit of 120 requests a minute,
-# since a request refused for a limit is never inspected. Fail if servewaf
-# refuses any of them.
-ordinary_requests() {
-    send /user/login --cookie-jar "$WORK/cookies" >/dev/null
-    status="$(send /user/login --cookie "$WORK/cookies" --cookie-jar "$WORK/cookies" \
-        --header "Origin: https://$SITE" --data-urlencode user_name=owner \
-        --data-urlencode "password=$PASSWORD")"
-    [ "$status" = 303 ] || fail "signing in was answered $status"
-
-    content="$(printf 'declare const env: Record<string, string>;\n' | base64)"
-    status="$(send /api/v1/repos/owner/repo/contents/process.env.d.ts \
-        --user "owner:$PASSWORD" --header "Content-Type: application/json" \
-        --data "{\"message\":\"Add process.env.d.ts\",\"content\":\"$content\"}")"
-    [ "$status" = 201 ] || fail "saving a file through the API was answered $status"
-    status="$(send /api/v1/repos/owner/repo/issues --user "owner:$PASSWORD" \
-        --header "Content-Type: application/json" \
-        --data '{"title":"ls -la fails","body":"SELECT * FROM users; <img src=x>"}')"
-    [ "$status" = 201 ] || fail "opening an issue through the API was answered $status"
-
-    grep -v -e '^#' -e '^$' "$ROOT/internal/waf/testdata/gitea-requests.txt" |
-        while read -r target header; do
-            if [ -n "$header" ]; then
-                send "$target" --globoff --cookie "$WORK/cookies" \
-                    --header "Referer: https://$SITE$target" --header "$header" >/dev/null
-            else
-                send "$target" --globoff --cookie "$WORK/cookies" \
-                    --header "Referer: https://$SITE$target" >/dev/null
-            fi
-            sleep 1
-        done
-
-    # The last request's report is logged once its answer has ended.
-    send /api/v1/settings/api >/dev/null
-    wait_for "servewaf logged no report of the last ordinary request" logged servewaf \
-        '"servewaf_line":"report"' '"path":"/api/v1/settings/api"'
-    refusals="$(refused)"
-    [ -z "$refusals" ] || fail "servewaf refused ordinary requests:
-$refusals"
-}
-
-# attacks: send SQL injection, path traversal and script injection in a query
-# string, and fail unless the Core Rule Set refuses each.
-attacks() {
-    # The traversal climbs one directory: the default rule file refuses two or
-    # more first, before the Core Rule Set runs.
-    for query in "id=1'%20OR%20'1'%3D'1" "id=..%2Fcustom%2Fconf%2Fapp.ini" \
-        "id=%3Cscript%3Ealert(1)%3C%2Fscript%3E"; do
-        status="$(send "/owner/repo?$query")"
-        [ "$status" = 403 ] || fail "the attack ?$query was answered $status"
-        wait_for "servewaf logged no refusal of ?$query" logged servewaf \
-            '"servewaf_line":"report"' '"action":"waf_blocked"' "\"query\":\"$query\""
-    done
-}
-
 main() {
     cd "$ROOT"
     trap cleanup EXIT
@@ -275,12 +204,6 @@ main() {
     esac
     echo "gitea-setup: a request from the shared network skips the checks"
 
-    site_git -C "$WORK/clone" fetch --quiet || fail "the fetch failed"
-    ordinary_requests
-    echo "gitea-setup: the Core Rule Set refuses no ordinary gitea request"
-    attacks
-    echo "gitea-setup: the Core Rule Set refuses the attacks"
-
     compose stop servewaf >/dev/null 2>&1 || fail "servewaf did not stop"
     status="$(send /owner/repo)"
     [ "$status" = 200 ] ||