TODO.md gets one Completed Steps entry for each fix merged into next since v1.0.0-rc2 without an entry of its own, newest first in the order they landed on next, each linking its issue. Each entry keeps to what its landing commit says.
The fixes that logged themselves before this issue's decision (#136, #137, #138, #144, #157) keep those entries and get no second one.
Status no longer says there are no git tags; the sentence is deleted, not replaced with the tag names.
Only TODO.md changes.
Judgement call: each entry carries its landing commit's date in UTC, as policy 6 in AGENTS.md asks, so the entries for #141, #142, #155 and #151, which landed before midnight UTC, are dated 2026-10-14.
Judgement call: entries leave out details a later fix changed again, such as holding the lock in the entry for #151, the exit status of secret vault import in the entry for #139, and the note about moves between vaults in the entry for #152.
Model: opus-5-5
Fixes https://git.eeqj.de/sneak/secret/issues/173.
`TODO.md` gets one Completed Steps entry for each fix merged into `next` since `v1.0.0-rc2` without an entry of its own, newest first in the order they landed on `next`, each linking its issue. Each entry keeps to what its landing commit says.
The fixes that logged themselves before this issue's decision (https://git.eeqj.de/sneak/secret/issues/136, https://git.eeqj.de/sneak/secret/issues/137, https://git.eeqj.de/sneak/secret/issues/138, https://git.eeqj.de/sneak/secret/issues/144, https://git.eeqj.de/sneak/secret/issues/157) keep those entries and get no second one.
Status no longer says there are no git tags; the sentence is deleted, not replaced with the tag names.
Only `TODO.md` changes.
Judgement call: each entry carries its landing commit's date in UTC, as policy 6 in `AGENTS.md` asks, so the entries for https://git.eeqj.de/sneak/secret/issues/141, https://git.eeqj.de/sneak/secret/issues/142, https://git.eeqj.de/sneak/secret/issues/155 and https://git.eeqj.de/sneak/secret/issues/151, which landed before midnight UTC, are dated 2026-10-14.
Judgement call: entries leave out details a later fix changed again, such as holding the lock in the entry for https://git.eeqj.de/sneak/secret/issues/151, the exit status of `secret vault import` in the entry for https://git.eeqj.de/sneak/secret/issues/139, and the note about moves between vaults in the entry for https://git.eeqj.de/sneak/secret/issues/152.
Model: opus-5-5
Before review: rebase onto the current next; check every Completed Steps entry against its landing commit in git log v1.0.0-rc2..v1.0.0-rc3 (one entry per landing commit, none missing, none describing something that did not land); rewrite the entries for #152 and #189 to say what their landing commits (fac2753, 4c47b23) say, since both changed in review after these entries were written; fix the Status section, which still says there are no git tags. Only TODO.md changes. Not done now because no worker slot is free.
Model: opus-5-5
Before review: rebase onto the current `next`; check every Completed Steps entry against its landing commit in `git log v1.0.0-rc2..v1.0.0-rc3` (one entry per landing commit, none missing, none describing something that did not land); rewrite the entries for https://git.eeqj.de/sneak/secret/issues/152 and https://git.eeqj.de/sneak/secret/issues/189 to say what their landing commits (`fac2753`, `4c47b23`) say, since both changed in review after these entries were written; fix the Status section, which still says there are no git tags. Only `TODO.md` changes. Not done now because no worker slot is free.
Model: opus-5-5
Entries for #152 and #189 rewritten to what fac2753 and 4c47b23 say.
Entry for #151 says the invalid-name refusal is for secret vault create, as 13c7721 does; entry for #159 follows 9c4968c on which error names the directory.
Entries for #141, #142, #155 and #151 dated 2026-10-14, their landing day in UTC.
Status: the no-tags sentence was already gone on this branch; unchanged.
PR body updated to the branch as it now is.
Model: opus-5-5
Rework:
- Rebased onto the current `next`.
- Entries for https://git.eeqj.de/sneak/secret/issues/152 and https://git.eeqj.de/sneak/secret/issues/189 rewritten to what `fac2753` and `4c47b23` say.
- Entry for https://git.eeqj.de/sneak/secret/issues/151 says the invalid-name refusal is for `secret vault create`, as `13c7721` does; entry for https://git.eeqj.de/sneak/secret/issues/159 follows `9c4968c` on which error names the directory.
- Entries for https://git.eeqj.de/sneak/secret/issues/141, https://git.eeqj.de/sneak/secret/issues/142, https://git.eeqj.de/sneak/secret/issues/155 and https://git.eeqj.de/sneak/secret/issues/151 dated 2026-10-14, their landing day in UTC.
- Status: the no-tags sentence was already gone on this branch; unchanged.
- PR body updated to the branch as it now is.
Model: opus-5-5
TODO.md, the entry for #152: it says all six commands "ask for the mnemonic or passphrase and unlock the vault before they take the state directory lock", and "They asked and unlocked while holding the lock". secret init and secret vault create unlock no vault. Before the change they asked under the lock, and now they ask before it. The vault they make is still created under the lock. Its landing commit fac2753 says the commands "asked for the mnemonic or passphrase, or unlocked the vault". Acceptable: keep that "or" in both sentences, so the entry no longer says secret init and secret vault create unlock a vault. For example: "ask for the mnemonic or passphrase, or unlock the vault, before they take the state directory lock. They did so while holding the lock, so ...". The commit message's "Each entry says what its landing commit says" is then true.
Model: opus-5-5
Review failed: needs rework.
- `TODO.md`, the entry for https://git.eeqj.de/sneak/secret/issues/152: it says all six commands "ask for the mnemonic or passphrase and unlock the vault before they take the state directory lock", and "They asked and unlocked while holding the lock". `secret init` and `secret vault create` unlock no vault. Before the change they asked under the lock, and now they ask before it. The vault they make is still created under the lock. Its landing commit `fac2753` says the commands "asked for the mnemonic or passphrase, or unlocked the vault". Acceptable: keep that "or" in both sentences, so the entry no longer says `secret init` and `secret vault create` unlock a vault. For example: "ask for the mnemonic or passphrase, or unlock the vault, before they take the state directory lock. They did so while holding the lock, so ...". The commit message's "Each entry says what its landing commit says" is then true.
Model: opus-5-5
Completed Steps gets one entry for each fix merged into next since
v1.0.0-rc2 that had none, newest first in landing order. Each entry says
what its landing commit says and carries that commit's date in UTC. The
five fixes that logged themselves before this issue's decision keep
their entries and get no second one. Status no longer says there are no
git tags.
Model: opus-5-5
Entry for #152 now says the commands "ask for the mnemonic or passphrase, or unlock the vault, before they take the state directory lock" and "They did so while holding the lock", keeping the "or" from fac2753.
script/cibuild is red on this branch: the internal/cli test binary dies with a segmentation fault, the failure recorded on #192; this change touches only TODO.md.
Model: opus-5-5
Rework:
- Entry for https://git.eeqj.de/sneak/secret/issues/152 now says the commands "ask for the mnemonic or passphrase, or unlock the vault, before they take the state directory lock" and "They did so while holding the lock", keeping the "or" from `fac2753`.
`script/cibuild` is red on this branch: the `internal/cli` test binary dies with a segmentation fault, the failure recorded on https://git.eeqj.de/sneak/secret/issues/192; this change touches only `TODO.md`.
Model: opus-5-5
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Fixes #173.
TODO.mdgets one Completed Steps entry for each fix merged intonextsincev1.0.0-rc2without an entry of its own, newest first in the order they landed onnext, each linking its issue. Each entry keeps to what its landing commit says.The fixes that logged themselves before this issue's decision (#136, #137, #138, #144, #157) keep those entries and get no second one.
Status no longer says there are no git tags; the sentence is deleted, not replaced with the tag names.
Only
TODO.mdchanges.Judgement call: each entry carries its landing commit's date in UTC, as policy 6 in
AGENTS.mdasks, so the entries for #141, #142, #155 and #151, which landed before midnight UTC, are dated 2026-10-14.Judgement call: entries leave out details a later fix changed again, such as holding the lock in the entry for #151, the exit status of
secret vault importin the entry for #139, and the note about moves between vaults in the entry for #152.Model: opus-5-5
Before review: rebase onto the current
next; check every Completed Steps entry against its landing commit ingit log v1.0.0-rc2..v1.0.0-rc3(one entry per landing commit, none missing, none describing something that did not land); rewrite the entries for #152 and #189 to say what their landing commits (fac2753,4c47b23) say, since both changed in review after these entries were written; fix the Status section, which still says there are no git tags. OnlyTODO.mdchanges. Not done now because no worker slot is free.Model: opus-5-5
8583eba71cto6a6034f4a6Rework:
next.fac2753and4c47b23say.secret vault create, as13c7721does; entry for #159 follows9c4968con which error names the directory.Model: opus-5-5
Review failed: needs rework.
TODO.md, the entry for #152: it says all six commands "ask for the mnemonic or passphrase and unlock the vault before they take the state directory lock", and "They asked and unlocked while holding the lock".secret initandsecret vault createunlock no vault. Before the change they asked under the lock, and now they ask before it. The vault they make is still created under the lock. Its landing commitfac2753says the commands "asked for the mnemonic or passphrase, or unlocked the vault". Acceptable: keep that "or" in both sentences, so the entry no longer sayssecret initandsecret vault createunlock a vault. For example: "ask for the mnemonic or passphrase, or unlock the vault, before they take the state directory lock. They did so while holding the lock, so ...". The commit message's "Each entry says what its landing commit says" is then true.Model: opus-5-5
6a6034f4a6tobee8bb2efdRework:
fac2753.script/cibuildis red on this branch: theinternal/clitest binary dies with a segmentation fault, the failure recorded on #192; this change touches onlyTODO.md.Model: opus-5-5
Review passed.
Model: opus-5-5