Carry the dated manifest's surviving newcomers into every packet (--manifest-newcomers) - #151

…a refuted discovery (Codex review of #151)

- LABEL_BEARING_RECEIPT also covers the manifest disposition vocabulary (refuted, keep_but_compare,
  targeted_candidate, newcomer, discovered, reject, demot, disposition); none of the 142 registered receipts
  matches the added terms.
- A refuted_* disposition withholds only a newcomer addition: the 2026-09-23 refutations of ledger
  repositories refute the discovery proposal ("not new to the catalog", "already conditional"), not the
  repository, so a ledger candidate keeps its place (documented and tested).

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…a refuted discovery (Codex review of #151)

- LABEL_BEARING_RECEIPT also covers the manifest disposition vocabulary (refuted, keep_but_compare,
  targeted_candidate, newcomer, discovered, reject, demot, disposition); none of the 142 registered receipts
  matches the added terms.
- A refuted_* disposition withholds only a newcomer addition: the 2026-09-23 refutations of ledger
  repositories refute the discovery proposal ("not new to the catalog", "already conditional"), not the
  repository, so a ledger candidate keeps its place (documented and tested).

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…ew of #151)

Hugging Face model candidates have no GitHub slug and were dropped. Under --manifest-newcomers a non-GitHub
https repository is identified by its lowercased URL (no trailing slash or .git); the default build keeps the
slug-only rule. On manifest-20260923 the blind build now carries 312 candidates, 6 of them Hugging Face models.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

… with (Codex review of #151)

An evidence[] entry such as '<path> items[x].corrections[0]' now attaches <path> (registered, unchanged,
not label-bearing when blind); the locator is dropped. The 2026-09-23 manifest names one such file; the
landscape-sweep source-review receipts supply the rest.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…a refuted discovery (Codex review of #151)

- LABEL_BEARING_RECEIPT also covers the manifest disposition vocabulary (refuted, keep_but_compare,
  targeted_candidate, newcomer, discovered, reject, demot, disposition); none of the 142 registered receipts
  matches the added terms.
- A refuted_* disposition withholds only a newcomer addition: the 2026-09-23 refutations of ledger
  repositories refute the discovery proposal ("not new to the catalog", "already conditional"), not the
  repository, so a ledger candidate keeps its place (documented and tested).

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…ew of #151)

Hugging Face model candidates have no GitHub slug and were dropped. Under --manifest-newcomers a non-GitHub
https repository is identified by its lowercased URL (no trailing slash or .git); the default build keeps the
slug-only rule. On manifest-20260923 the blind build now carries 312 candidates, 6 of them Hugging Face models.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

… with (Codex review of #151)

An evidence[] entry such as '<path> items[x].corrections[0]' now attaches <path> (registered, unchanged,
not label-bearing when blind); the locator is dropped. The 2026-09-23 manifest names one such file; the
landscape-sweep source-review receipts supply the rest.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…a refuted discovery (Codex review of #151)

- LABEL_BEARING_RECEIPT also covers the manifest disposition vocabulary (refuted, keep_but_compare,
  targeted_candidate, newcomer, discovered, reject, demot, disposition); none of the 142 registered receipts
  matches the added terms.
- A refuted_* disposition withholds only a newcomer addition: the 2026-09-23 refutations of ledger
  repositories refute the discovery proposal ("not new to the catalog", "already conditional"), not the
  repository, so a ledger candidate keeps its place (documented and tested).

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…ew of #151)

Hugging Face model candidates have no GitHub slug and were dropped. Under --manifest-newcomers a non-GitHub
https repository is identified by its lowercased URL (no trailing slash or .git); the default build keeps the
slug-only rule. On manifest-20260923 the blind build now carries 312 candidates, 6 of them Hugging Face models.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

… with (Codex review of #151)

An evidence[] entry such as '<path> items[x].corrections[0]' now attaches <path> (registered, unchanged,
not label-bearing when blind); the locator is dropped. The 2026-09-23 manifest names one such file; the
landscape-sweep source-review receipts supply the rest.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…a refuted discovery (Codex review of #151)

- LABEL_BEARING_RECEIPT also covers the manifest disposition vocabulary (refuted, keep_but_compare,
  targeted_candidate, newcomer, discovered, reject, demot, disposition); none of the 142 registered receipts
  matches the added terms.
- A refuted_* disposition withholds only a newcomer addition: the 2026-09-23 refutations of ledger
  repositories refute the discovery proposal ("not new to the catalog", "already conditional"), not the
  repository, so a ledger candidate keeps its place (documented and tested).

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…ew of #151)

Hugging Face model candidates have no GitHub slug and were dropped. Under --manifest-newcomers a non-GitHub
https repository is identified by its lowercased URL (no trailing slash or .git); the default build keeps the
slug-only rule. On manifest-20260923 the blind build now carries 312 candidates, 6 of them Hugging Face models.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

… with (Codex review of #151)

An evidence[] entry such as '<path> items[x].corrections[0]' now attaches <path> (registered, unchanged,
not label-bearing when blind); the locator is dropped. The 2026-09-23 manifest names one such file; the
landscape-sweep source-review receipts supply the rest.

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

…ory assistant capture opt-in, ignore Claude Code local state (#159)

* Adoption hygiene: native Claude Code pin is a floor (2.1.281), ai-memory assistant capture is opt-in, ignore Claude Code local state

bootstrap-macos.sh's install_native ran `"$bin" install <pin>` even when Claude Code's
native auto-updater had already installed something newer, so re-running the bootstrap
downgraded it. It now keeps a ~/.local/bin/<bin> whose --version is at or above the pin
(logs "Kept", downloads and installs nothing); a missing, older or unreadable launcher
takes the unchanged checksum-verified install. The check lives inside install_native
because tests/test_adoption_bootstrap.py runs that function on its own. Bash 3.2 safe.

Both claude-code pins move to 2.1.281 together (shared components keep one version on
both platforms). darwin-arm64 a922981f... comes from the official 2.1.281 manifest.json
and matches, byte for byte, the copy the native updater installed on a Mac. linux-x64
56fe3da8... comes from the same manifest and was not re-hashed; fetch() checks it before
running anything. 2.1.281 fixes a recursive rm of command-substitution output running
unprompted in auto and --dangerously-skip-permissions mode. The macOS page's pin table
and the generated new-host grand list follow the pins.

The Claude settings template no longer passes --capture-assistant to ai-memory's Stop
hook. render_config.py adds it only when AI_MEMORY_CAPTURE_ASSISTANT=true (host value
file or --set) and rejects any value other than true, false or empty. defaultMode and
skipDangerousModePermissionPrompt are unchanged.

.gitignore adds .claude/settings.local.json and .claude/worktrees/.

Not included, because open PRs #116, #145 and #151 own these files: the
manifests/evidence.json re-hash of the 12 changed files, and the adoption/bootstrap.md
"changed after `v2026.09.23.1`" marker for the three changed install inputs.

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

* Re-hash evidence and mark the Claude Code floor in the bootstrap guide

Records the SHA-256/byte entries for every file this branch changes and
adds the "changed after `v2026.09.23.1`" note that the docs-consistency
test requires for the new install behavior (pins 2.1.281, floor instead
of unconditional reinstall). Both files are also edited by open PRs;
whichever lands later re-runs the same host_receipts.register_file helper.

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

---------

Co-authored-by: seathatflowsinourveins <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Scout <scout@local>

…loor (stacked on #159) (#177)

* Adoption hygiene: native Claude Code pin is a floor (2.1.281), ai-memory assistant capture is opt-in, ignore Claude Code local state

bootstrap-macos.sh's install_native ran `"$bin" install <pin>` even when Claude Code's
native auto-updater had already installed something newer, so re-running the bootstrap
downgraded it. It now keeps a ~/.local/bin/<bin> whose --version is at or above the pin
(logs "Kept", downloads and installs nothing); a missing, older or unreadable launcher
takes the unchanged checksum-verified install. The check lives inside install_native
because tests/test_adoption_bootstrap.py runs that function on its own. Bash 3.2 safe.

Both claude-code pins move to 2.1.281 together (shared components keep one version on
both platforms). darwin-arm64 a922981f... comes from the official 2.1.281 manifest.json
and matches, byte for byte, the copy the native updater installed on a Mac. linux-x64
56fe3da8... comes from the same manifest and was not re-hashed; fetch() checks it before
running anything. 2.1.281 fixes a recursive rm of command-substitution output running
unprompted in auto and --dangerously-skip-permissions mode. The macOS page's pin table
and the generated new-host grand list follow the pins.

The Claude settings template no longer passes --capture-assistant to ai-memory's Stop
hook. render_config.py adds it only when AI_MEMORY_CAPTURE_ASSISTANT=true (host value
file or --set) and rejects any value other than true, false or empty. defaultMode and
skipDangerousModePermissionPrompt are unchanged.

.gitignore adds .claude/settings.local.json and .claude/worktrees/.

Not included, because open PRs #116, #145 and #151 own these files: the
manifests/evidence.json re-hash of the 12 changed files, and the adoption/bootstrap.md
"changed after `v2026.09.23.1`" marker for the three changed install inputs.

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

* Re-hash evidence and mark the Claude Code floor in the bootstrap guide

Records the SHA-256/byte entries for every file this branch changes and
adds the "changed after `v2026.09.23.1`" note that the docs-consistency
test requires for the new install behavior (pins 2.1.281, floor instead
of unconditional reinstall). Both files are also edited by open PRs;
whichever lands later re-runs the same host_receipts.register_file helper.

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

* Keep a newer native Claude Code on Linux/WSL2: the pin is a floor there too

adoption/bootstrap-linux.sh's install_native ran `"$bin" install <pin>`
unconditionally, so on a Linux/WSL2 host whose native updater (latest channel)
was ahead of the pin, re-running the bootstrap moved Claude Code back to the
pin and dropped newer fixes. install_native is now the same function as
bootstrap-macos.sh's (#159): a ~/.local/bin/<bin> whose --version first word
is a dotted numeric version at or above the pin is kept (nothing downloaded
or installed; install_pin logs "Kept installed ..." instead of "Installed
..."); a missing, older, non-numeric or failing launcher takes the unchanged
checksum-verified install.

Tests: NativeInstallFloorTests in tests/test_adoption_bootstrap.py run the
verbatim fetch/install_native/install_pin against a one-pin fixture (curl
shimmed; a shasum-backed sha256sum shim only where GNU --check --status is
missing), plus an identity check that both scripts share one install_native.

Docs: the Linux claude-code install_note describes the floor; bootstrap.md
step 2's stale "Linux/WSL2's script and pins are the same at v2026.09.23 as
on main" now says what differs at that tag; the claude-code floor notes on
bootstrap.md, linux-wsl2.md and macos-arm64.md name the currently pinned
release v2026.09.24.1 (#171), which the docs-consistency test requires.

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

* linux-wsl2.md: keep the changed-after marker on one line (docs-consistency)

Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>

---------

Co-authored-by: seathatflowsinourveins <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Co-authored-by: Scout <scout@local>

…heir approved text

The regression added for PR #985 (c9155c3) checked that the docs row and the
inventory entry name the research and Jev uses and the runner, and rejected the
fragments "separately authorized", "needs its own" and "separate key". The GPT
review of c9155c3 (CHANGES_REQUESTED, one P2 at
tests/test_credential_status.py:294, :295 and :299-301) measured three
contradictions of the owner's 2026-10-10 grant that all passed it, rc 0:
  - docs/secret-storage.md:20, the dated grant clause replaced by
    ", excluding the trading lane's research work and Jev harness"
  - "A trading-lane use requires its own independently authorized key."
    appended to the docs row
  - the same sentence appended to the inventory notes
    (adoption/credential-inventory.json:140)

test_typesafe_row_keeps_the_research_and_jev_grant_without_a_separate_key keeps
its name: it still says what it enforces. In both real files it now requires
the dated affirmative grant and the complete approved text after whitespace
normalisation (the docs row and the inventory notes), and loaders equal to
exactly the runner command. The phrase blacklist is gone. The same check runs
on planted contradictions that must fail it with the expected message: the
grant revoked, and each of two key-requirement sentences appended (the one
8e9e56c removed and the reworded one), in both places. A whitespace-only
reflow must still pass, and the docs row is found after whitespace
normalisation so a reformatted table does not hide it.

tests/test_credential_status.py is re-registered in manifests/evidence.json
with host_receipts.register_file (docs/lanes.md:107); only its sha256 and
bytes change.

Corrections audit (#151)
  Source: the GPT verdict for PR #985 at c9155c3, ~/.cache/ns2604-gpt-reads/
  fix2-pr985-c9155c3d8/VERDICT.md (sha256
  ccd7bffc9115c3770a481e2eef1b4b663f34ff2e6efe4257f0f7b7c60cb52ba6) with its
  REVIEW-EVIDENCE.md; the repo's own docs/secret-storage.md:20 and
  adoption/credential-inventory.json:140 at c9155c3. Mechanism, Python
  3.13.16 standard library as installed: unittest/case.py:1324-1338
  (assertRaisesRegex), :746-765 (_formatMessage appends the explicit message to
  the standard one after " : ", which the planted-case regexes rely on) and
  :537-566 (subTest).
  Fail before: the c9155c3 test returns rc 0 on each of the three
  contradictions above.
  Fail after: this test returns rc 1 on each (docs revocation: "docs/secret-
  storage.md row lost the dated grant"; the two appended sentences: "departs
  from the approved text"), and on the exact 2a33680 docs file (194,754
  bytes, sha256 028f69a6...), the dropped grant and runner in either file, and
  a revoked inventory grant. Each run is the single-test command with one real
  input file changed in an isolated archive copy of c9155c3.
  Pass after: the single-test command rc 0, 1 test; tests.test_credential_status
  rc 0, 49 tests; scripts/validate.py rc 0 (11421 hashed files);
  scripts/evidence_manifest.py --check rc 0. Whitespace-only changes to the
  docs row and the notes also pass, rc 0.
  Untested boundary: text outside the pinned docs row, notes and loaders, such
  as the inventory rotation field, is not bound (a sentence appended there
  still passes, rc 0).

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

… Jev harness (#985)

* Authorize the stored TypeSafe key for the trading lane's research and Jev harness

The owner decided on 2026-10-10 (north-star plan, owner item 2) that the trading
lane's research work and Jev harness may use the stored typesafe key instead of a
separate one. Each use still runs in its own process started through the
credential runner, and no lane environment holds the key. The inventory note and
the secret-storage row now record that; the lane field keeps its single-string
schema. Both files are re-registered in manifests/evidence.json per the hot-file
protocol.

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

* docs: drop the stale separate-key sentence from the typesafe row

The GPT read of #985 found docs/secret-storage.md:20 still said a trading-lane use needs its own,
separately authorized key, contradicting this PR's inventory note and body. The sentence is
removed; the row keeps the narrow research and Jev-harness grant and the credential-runner
requirement. Also restores the missing list comma. Re-registered per the hot-file protocol.

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

* test(credentials): pin the TypeSafe row's grant and runner command

PR #985 corrected docs/secret-storage.md:20 so the typesafe row agrees with
adoption/credential-inventory.json (owner decision 2026-10-10: the trading
lane's research work and Jev harness may also use the stored key, each in its
own process started through credential_run.py). No test caught the stale
sentence "A trading-lane use needs its own, separately authorized key", and
the GPT range-diff read of 8e9e56c asked for a committed regression.

The new test in tests/test_credential_status.py,
test_typesafe_row_keeps_the_research_and_jev_grant_without_a_separate_key,
asserts that the inventory row keeps the runner loader and the grant, that
the one docs row keeps the runner command and the grant, and that neither
carries a separate-key requirement.

tests/test_credential_status.py is re-registered in manifests/evidence.json
with host_receipts.register_file (docs/lanes.md:107); only its sha256 and
bytes change. scripts/evidence_manifest.py only sorts and checks the order
of files[] and has no hash-update mode.

Verdict: ~/.cache/ns2604-gpt-reads/fix1b-pr985-8e9e56cdc/VERDICT.md
(sha256 b057e6eedb496ae228d1a54805e3814da690e77494b53c53c961092950fccc03).

Fail before, with docs/secret-storage.md checked out from 2a33680 and
restored afterwards (adoption/credential-inventory.json is identical at both):
  git checkout 2a33680 -- docs/secret-storage.md
  python3 -B -m unittest tests.test_credential_status.CredentialStatusTests.test_typesafe_row_keeps_the_research_and_jev_grant_without_a_separate_key
    rc 1, 1 test, failures=1: 'separately authorized' unexpectedly found
    (place='docs/secret-storage.md row')
  python3 -B -m unittest tests.test_credential_status
    rc 1, 49 tests, failures=1 (only this test)

Pass after, at this commit:
  the same single-test command: rc 0, 1 test
  python3 -B -m unittest tests.test_credential_status: rc 0, 49 tests (48 before)
  python3 scripts/validate.py: rc 0 (11421 hashed files)
  python3 scripts/evidence_manifest.py --check: rc 0

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

* test(credentials): pin the TypeSafe docs row and inventory notes to their approved text

The regression added for PR #985 (c9155c3) checked that the docs row and the
inventory entry name the research and Jev uses and the runner, and rejected the
fragments "separately authorized", "needs its own" and "separate key". The GPT
review of c9155c3 (CHANGES_REQUESTED, one P2 at
tests/test_credential_status.py:294, :295 and :299-301) measured three
contradictions of the owner's 2026-10-10 grant that all passed it, rc 0:
  - docs/secret-storage.md:20, the dated grant clause replaced by
    ", excluding the trading lane's research work and Jev harness"
  - "A trading-lane use requires its own independently authorized key."
    appended to the docs row
  - the same sentence appended to the inventory notes
    (adoption/credential-inventory.json:140)

test_typesafe_row_keeps_the_research_and_jev_grant_without_a_separate_key keeps
its name: it still says what it enforces. In both real files it now requires
the dated affirmative grant and the complete approved text after whitespace
normalisation (the docs row and the inventory notes), and loaders equal to
exactly the runner command. The phrase blacklist is gone. The same check runs
on planted contradictions that must fail it with the expected message: the
grant revoked, and each of two key-requirement sentences appended (the one
8e9e56c removed and the reworded one), in both places. A whitespace-only
reflow must still pass, and the docs row is found after whitespace
normalisation so a reformatted table does not hide it.

tests/test_credential_status.py is re-registered in manifests/evidence.json
with host_receipts.register_file (docs/lanes.md:107); only its sha256 and
bytes change.

Corrections audit (#151)
  Source: the GPT verdict for PR #985 at c9155c3, ~/.cache/ns2604-gpt-reads/
  fix2-pr985-c9155c3d8/VERDICT.md (sha256
  ccd7bffc9115c3770a481e2eef1b4b663f34ff2e6efe4257f0f7b7c60cb52ba6) with its
  REVIEW-EVIDENCE.md; the repo's own docs/secret-storage.md:20 and
  adoption/credential-inventory.json:140 at c9155c3. Mechanism, Python
  3.13.16 standard library as installed: unittest/case.py:1324-1338
  (assertRaisesRegex), :746-765 (_formatMessage appends the explicit message to
  the standard one after " : ", which the planted-case regexes rely on) and
  :537-566 (subTest).
  Fail before: the c9155c3 test returns rc 0 on each of the three
  contradictions above.
  Fail after: this test returns rc 1 on each (docs revocation: "docs/secret-
  storage.md row lost the dated grant"; the two appended sentences: "departs
  from the approved text"), and on the exact 2a33680 docs file (194,754
  bytes, sha256 028f69a6...), the dropped grant and runner in either file, and
  a revoked inventory grant. Each run is the single-test command with one real
  input file changed in an isolated archive copy of c9155c3.
  Pass after: the single-test command rc 0, 1 test; tests.test_credential_status
  rc 0, 49 tests; scripts/validate.py rc 0 (11421 hashed files);
  scripts/evidence_manifest.py --check rc 0. Whitespace-only changes to the
  docs row and the notes also pass, rc 0.
  Untested boundary: text outside the pinned docs row, notes and loaders, such
  as the inventory rotation field, is not bound (a sentence appended there
  still passes, rc 0).

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

---------

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