Authenticate CI Claude review with an API key in a main-only environment - #968

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

… git

The GPT read at 13d1eb7 asked, under the correction rule, for:
- the upstream sources of the environment reading: GitHub's workflow syntax
  for jobs.<job_id>.environment (a name, or an object with name and url) and
  YAML 1.2.2's block and flow mappings, quoted keys and comments, cited with
  page digests in the record and next to the regression cases;
- a regression for the corrected inverse pin: RetirementInversePinTests reads
  the pin from the record and checks it is #968's merge commit holding the
  three workflows. It fails on the former SHA (no such commit) and is skipped,
  with the reason, only outside a worktree or in a shallow clone (CI clones
  with fetch-depth 0).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

… git

The GPT read at 13d1eb7 asked, under the correction rule, for:
- the upstream sources of the environment reading: GitHub's workflow syntax
  for jobs.<job_id>.environment (a name, or an object with name and url) and
  YAML 1.2.2's block and flow mappings, quoted keys and comments, cited with
  page digests in the record and next to the regression cases;
- a regression for the corrected inverse pin: RetirementInversePinTests reads
  the pin from the record and checks it is #968's merge commit holding the
  three workflows. It fails on the former SHA (no such commit) and is skipped,
  with the reason, only outside a worktree or in a shallow clone (CI clones
  with fetch-depth 0).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

… git

The GPT read at 13d1eb7 asked, under the correction rule, for:
- the upstream sources of the environment reading: GitHub's workflow syntax
  for jobs.<job_id>.environment (a name, or an object with name and url) and
  YAML 1.2.2's block and flow mappings, quoted keys and comments, cited with
  page digests in the record and next to the regression cases;
- a regression for the corrected inverse pin: RetirementInversePinTests reads
  the pin from the record and checks it is #968's merge commit holding the
  three workflows. It fails on the former SHA (no such commit) and is skipped,
  with the reason, only outside a worktree or in a shallow clone (CI clones
  with fetch-depth 0).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Retire Claude review inside GitHub; AI reviews run locally

The owner decided on 2026-10-10 (about 17:15Z, relayed by the command center)
that Claude review no longer runs inside GitHub: AI reviews run on the local
host and GitHub enforces the gates. The API refused every request that carried
claude-code-action's attribution (CLAUDE_CODE_ENTRYPOINT=claude-code-github-action,
base-action/src/parse-sdk-options.ts:284 at 2dca132f) on all five Console keys,
while the same keys work from the CLI and the Agent SDK (the command center's
correction #141, measured).

Removed: claude-pr-review.yml, claude-pr-toolkit-review.yml, harness-audit.yml
and their three test modules, their zizmor-coverage entries and their
local-pages inventory paths. test_claude_federation_diagnostics keeps its
retained-snapshot and decision-claim tests. A new
InGitHubClaudeReviewRetiredTests keeps the three files absent and no workflow
running the Claude action or naming an Anthropic key.

New record docs/decisions/2026-10-10-retire-in-github-claude-review.md (cause,
what remains, revisit triggers, inverse); supersession notes in the four
claude-actions/federation records and the decision-record index;
docs/github-automation.md's current practice.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Pin the repository grant count at 90 after the retirement

The three removed workflows were in the local-pages repository grants. At the
landing base, main f38116f (93 grants), the test's own derivation
(grants.inventory_paths over git ls-files) gives 90 for this branch, equal to
the policy's repository grants, and the test pins 90. At the earlier bases
eb31d23 and 9d4c00c the same derivation gave 90 and 89.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Take up the command center's notes on the retirement

- .github/actions-permissions.json drops anthropics/claude-code-action@* and
  oven-sh/setup-bun@*: no workflow uses them; a new test keeps every allowed
  pattern in use, and the record's inverse puts them back.
- docs/github-automation.md: the allow-list text, no "harness description",
  no "still awaits hosted acceptance" clause.
- The retirement test catches the claude-review environment in its plain,
  name: and quoted forms, and reads *.yaml as well.
- Record nits: #953 marked open, the check is named sota-sources, and the
  federation-diagnostics note says the three workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Cite #968's real merge commit in the inverse

The record's inverse named a SHA that does not exist; #968's merge commit is
9d4c00c (git rev-parse 9d4c00c). The
federation note now says the history of those workflows.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Read each job's environment in the retirement guard

The GPT read at 48e4bd3 found three valid YAML forms of the claude-review
environment that the regex guard let through: a flow mapping, a quoted name
key, and a comment after the key. The guard now reads each job's environment
as a scalar or a mapping's name (job_environments), cross-checked against
PyYAML when it is installed; all three forms are regressions, each caught here
and missed at e46f68e, and actionlint accepts each fixture.

The record names the claude-native-practice skill rows as a dated follow-up
for cc-native-practice (the command center's ruling at 20:19Z).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Word the skill-row follow-up without the pinned-copy premise

The command center's correction #146: no personal copy of claude-native-practice
is installed on this host; sessions load the checkout's project skill. The
follow-up is now stated neutrally: the skill rows for the retired route are
updated in a cc-native-practice pull request, then re-pinned.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* State what the environment reading covers, and what PyYAML covers

The docstring claimed comments in general; the text reading handles a comment
after the key, not a comment line before name: at the job's indentation or a
multi-line flow mapping. The PyYAML comparison covers those, and CI runs it
(validate run 38084944806, passed, not skipped). The command center's micro at
08a76ab.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Cite the environment forms' sources, and test the inverse pin against git

The GPT read at 13d1eb7 asked, under the correction rule, for:
- the upstream sources of the environment reading: GitHub's workflow syntax
  for jobs.<job_id>.environment (a name, or an object with name and url) and
  YAML 1.2.2's block and flow mappings, quoted keys and comments, cited with
  page digests in the record and next to the regression cases;
- a regression for the corrected inverse pin: RetirementInversePinTests reads
  the pin from the record and checks it is #968's merge commit holding the
  three workflows. It fails on the former SHA (no such commit) and is skipped,
  with the reason, only outside a worktree or in a shallow clone (CI clones
  with fetch-depth 0).

Co-Authored-By: Claude Opus 5.5 <[email protected]>

---------

Co-authored-by: Claude Opus 5.5 <[email protected]>

docs/secret-storage.md gains rows for anthropic-api-5 to anthropic-api-10
and a "Key practice (2026-10-10)" section, the repository copy of the
command center's record. It covers slot purposes by organization id prefix,
one purpose per slot, hidden-prompt storage, per-command use, main-only
environment secrets for CI, the health check, tracking, rotation, spend
control, and fill-first failover with anthropic-api-3 kept for CI. Owner
direction is attributed, not quoted; account names and balances are left
out.

The no-Admin-API and cache-isolation statements quote Anthropic's Admin
API, Usage and Cost API and prompt-caching pages, read 2026-10-10. The
gh stdin statement quotes the installed gh 2.102.0 help.

The inventory gives anthropic-api-3 its purpose (CI Claude review, the
main-only claude-review environment secret) and a rotation step that resets
that secret. The notes of anthropic-api-3 and anthropic-api-4 mark "no
Anthropic key is stored in GitHub" as superseded under the owner's ruling
of 2026-10-10 (PR #968). No other entry changes.

The new decision record names the evidence, alternatives, overturn
conditions and SOTA sources. As in #930, the harness-defaults correction row
now covers every slot, and the five secret-storage line citations in
upstream-surface-dispositions.json move by the 122 inserted lines.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

docs/secret-storage.md gains rows for anthropic-api-5 to anthropic-api-10
and a "Key practice (2026-10-10)" section, the repository copy of the
command center's record. It covers slot purposes by organization id prefix,
one purpose per slot, hidden-prompt storage, per-command use, main-only
environment secrets for CI, the health check, tracking, rotation, spend
control, and fill-first failover with anthropic-api-3 kept for CI. Owner
direction is attributed, not quoted; account names and balances are left
out.

The no-Admin-API and cache-isolation statements quote Anthropic's Admin
API, Usage and Cost API and prompt-caching pages, read 2026-10-10. The
gh stdin statement quotes the installed gh 2.102.0 help.

The inventory gives anthropic-api-3 its purpose (CI Claude review, the
main-only claude-review environment secret) and a rotation step that resets
that secret. The notes of anthropic-api-3 and anthropic-api-4 mark "no
Anthropic key is stored in GitHub" as superseded under the owner's ruling
of 2026-10-10 (PR #968). No other entry changes.

The new decision record names the evidence, alternatives, overturn
conditions and SOTA sources. As in #930, the harness-defaults correction row
now covers every slot, and the five secret-storage line citations in
upstream-surface-dispositions.json move by the 122 inserted lines.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

docs/secret-storage.md gains rows for anthropic-api-5 to anthropic-api-10
and a "Key practice (2026-10-10)" section, the repository copy of the
command center's record. It covers slot purposes by organization id prefix,
one purpose per slot, hidden-prompt storage, per-command use, main-only
environment secrets for CI, the health check, tracking, rotation, spend
control, and fill-first failover with anthropic-api-3 kept for CI. Owner
direction is attributed, not quoted; account names and balances are left
out.

The no-Admin-API and cache-isolation statements quote Anthropic's Admin
API, Usage and Cost API and prompt-caching pages, read 2026-10-10. The
gh stdin statement quotes the installed gh 2.102.0 help.

The inventory gives anthropic-api-3 its purpose (CI Claude review, the
main-only claude-review environment secret) and a rotation step that resets
that secret. The notes of anthropic-api-3 and anthropic-api-4 mark "no
Anthropic key is stored in GitHub" as superseded under the owner's ruling
of 2026-10-10 (PR #968). No other entry changes.

The new decision record names the evidence, alternatives, overturn
conditions and SOTA sources. As in #930, the harness-defaults correction row
now covers every slot, and the five secret-storage line citations in
upstream-surface-dispositions.json move by the 122 inserted lines.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

…ken health check (#969)

* Add Anthropic API key slots anthropic-api-5 to anthropic-api-10

Stored through the credential terminal on the owner's direction (2026-10-10).
No admin-key slots: every organization behind the stored keys is an individual
account, and the Admin API is unavailable for individual accounts (Anthropic docs,
read 2026-10-10). Data-only inventory entries; api-actions' PR adds the docs,
evidence and tests.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Cover anthropic-api-5 to anthropic-api-10 in the runner tests

The six slots added in the previous commit were not in INJECTABLE_IDS, so
both inventory-driven set assertions failed at that commit. ANTHROPIC_IDS
now lists all ten Anthropic slots; INJECTABLE_IDS and the same-grammar
setter/runner test use it, so every slot is written through the hidden-prompt
setter and injected through credential_run.py with synthetic values only.

Failing-first: with the e1c88ff inventory the three affected tests fail
(failures=14, errors=18: the setter and the runner refuse anthropic-api-5..10
as unknown ids, and both id sets differ); with this branch's inventory they
pass. The previous commit changed adoption/credential-inventory.json without
its evidence binding, so this commit registers it with the test file.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Add a no-token health check for the Anthropic key slots

tools/credentials/anthropic_key_health.py ports the command center's
anthropic_key_health.sh to the repository. For each anthropic-api slot
in inventory order it runs credential_run.py <slot> --check (starts
nothing), and for a stored slot runs the probe child through
credential_run.py <slot> --only ANTHROPIC_API_KEY. The key reaches only
the child's environment; the request headers go to the child as JSON on
stdin, so no argv carries a value. The child sends GET /v1/models and
GET /v1/organizations/me to the fixed API host, follows no redirect,
reads no body, and prints statuses and the first eight characters of
the anthropic-organization-id response header. The report never holds
a key, an organization name or a response body. The port drops the
script's organization-name read and its admin-key probe.

HTTP handling follows tools/credentials/alpaca_rate_limit_probe.py; the
header and endpoint are documented in the Claude API overview and the
Admin API page (read 2026-10-10). Tests use a fake opener, a fake
runner, and the real credential_run.py over a temporary store with a
synthetic key; no test makes a network call.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Record the Anthropic key slots and the 2026-10-10 key practice

docs/secret-storage.md gains rows for anthropic-api-5 to anthropic-api-10
and a "Key practice (2026-10-10)" section, the repository copy of the
command center's record. It covers slot purposes by organization id prefix,
one purpose per slot, hidden-prompt storage, per-command use, main-only
environment secrets for CI, the health check, tracking, rotation, spend
control, and fill-first failover with anthropic-api-3 kept for CI. Owner
direction is attributed, not quoted; account names and balances are left
out.

The no-Admin-API and cache-isolation statements quote Anthropic's Admin
API, Usage and Cost API and prompt-caching pages, read 2026-10-10. The
gh stdin statement quotes the installed gh 2.102.0 help.

The inventory gives anthropic-api-3 its purpose (CI Claude review, the
main-only claude-review environment secret) and a rotation step that resets
that secret. The notes of anthropic-api-3 and anthropic-api-4 mark "no
Anthropic key is stored in GitHub" as superseded under the owner's ruling
of 2026-10-10 (PR #968). No other entry changes.

The new decision record names the evidence, alternatives, overturn
conditions and SOTA sources. As in #930, the harness-defaults correction row
now covers every slot, and the five secret-storage line citations in
upstream-surface-dispositions.json move by the 122 inserted lines.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Apply the owner's revised fill-first order and anthropic-api-3's purpose

The owner revised the fill-first order at about 15:12Z on 2026-10-10
(relayed by the command center). The order is now anthropic-api-4,
anthropic-api-3, anthropic-api-5, anthropic-api, anthropic-api-2, then
anthropic-api-6 to anthropic-api-10 as stored. anthropic-api-3 serves the CI
Claude review first and local quality work second; it is no longer kept
out of the order. Its key remains the main-only claude-review environment
secret.

The key practice section, the practice table (anthropic-api-5 now follows
anthropic-api-3), the credentials table row, the inventory label and notes
of anthropic-api-3, and the decision record follow the revision. The
decision record also adds an overturn condition for CI credit drawn down
by local work, and corrects the CPython citation range to L623-L656. The
five secret-storage line citations in upstream-surface-dispositions.json
follow the doc's line count.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Rewrap two key-practice paragraphs

Two paragraphs of the key practice section carried uneven line breaks
from earlier edits; the text is unchanged. The rewrap removes one line,
so the five secret-storage line citations in
upstream-surface-dispositions.json move up by one.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Cite the refreshed key-practice record

The command center refreshed its record to the 15:12Z fill-first order; the
decision cites the new sha256.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

* Check the inventory first, and fail a probe that failed

The co-op's GPT read at 445c61c (2 P2):

- slots() now applies scripts/credential_status.py's inventory check,
  as the credential runner does, and refuses an anthropic-api entry in
  another shape or an inventory with none: exit 1 before any slot is
  checked, naming counts and inventory ids only. Before, entries: [] or
  an entry with no variables gave an empty report and exit 0.
- probe_slot() fails a child that exited nonzero or was killed
  (CompletedProcess.returncode, CPython 3.13.16 Lib/subprocess.py:484)
  before reading its output, and a line that reports an error is
  probe_failed whatever status it printed. Before, a 200 line followed
  by exit 7, a signal, or an error passed with exit 0.
- docs/secret-storage.md says so; the catalog's five citations of its
  later lines move by the four lines added.

Co-Authored-By: Claude Opus 5.5 <[email protected]>

---------

Co-authored-by: Claude Opus 5.5 <[email protected]>