fix(redact): exempt git identity vars from ENV-assignment redaction by Kyzcreig · Pull Request #3 · ANG-Ventures/hermes-agent

Kyzcreig · 2026-06-03T01:47:36Z

Problem

agent/redact.py's ENV-assignment redaction matches any VAR=value whose name contains a secret-like substring. GIT_AUTHOR_NAME and GIT_AUTHOR_EMAIL contain AUTH (inside AUTHOR), so they were falsely redacted to VAR=*** — mangling git commit-authoring commands in tool output.

This surfaced while re-authoring commits to satisfy the contributor-check attribution CI: every export GIT_AUTHOR_EMAIL=... got rewritten to ***, corrupting the author identity.

Fix

Add a leading word-boundary + negative-lookahead allowlist (_GIT_IDENTITY_ALLOWLIST) that exempts the four git identity vars (GIT_AUTHOR_NAME, GIT_AUTHOR_EMAIL, GIT_COMMITTER_NAME, GIT_COMMITTER_EMAIL) from ENV-assignment redaction. They are never secrets.

Verification (empirical, not just reasoning)

Git identity vars pass through untouched — including with an export prefix.
Every genuine secret STILL redacts: OPENAI_API_KEY, AWS_SECRET_ACCESS_KEY, *_AUTH_TOKEN, *_AUTHORIZATION_KEY.
3 new allowlist tests + 2 still-redact guards.
Full redact suite: 79 passed, 0 regressions (plus config-bridge + PII redaction suites green).

Risk

Minimal. Narrow allowlist of 4 well-known non-secret variable names; the secret-name matching is otherwise unchanged.

…redaction The ENV-assignment redaction pattern matches any VAR=value whose name contains a secret-like substring. GIT_AUTHOR_NAME and GIT_AUTHOR_EMAIL contain "AUTH" (inside "AUTHOR"), so they were being falsely redacted to VAR=*** mangling git commit-authoring commands in tool output (observed when re-authoring commits for the contributor-attribution CI check). Add a leading word-boundary + negative-lookahead allowlist (_GIT_IDENTITY_ALLOWLIST) that exempts the four git identity vars while still redacting every genuine secret. Verified: real keys (OPENAI_API_KEY, AWS_SECRET_ACCESS_KEY, *_AUTH_TOKEN, *_AUTHORIZATION_KEY) still redact; git identity vars (incl. with 'export ' prefix) pass through untouched. Tests: 3 new allowlist tests + 2 still-redact guards; full redact suite 79 passed, 0 regressions.

github-actions · 2026-06-03T01:48:24Z

🔎 Lint report: `fix/redact-git-identity-allowlist` vs `origin/main`

ruff

Total: 0 on HEAD, 0 on base (➖ 0)

🆕 New issues: none

✅ Fixed issues: none

Unchanged: 0 pre-existing issues carried over.

ty (type checker)

Total: 9592 on HEAD, 9592 on base (➖ 0)

🆕 New issues: none

✅ Fixed issues: none

Unchanged: 5059 pre-existing issues carried over.

Diagnostics are surfaced as warnings — this check never fails the build.

…redaction (#3) The ENV-assignment redaction pattern matches any VAR=value whose name contains a secret-like substring. GIT_AUTHOR_NAME and GIT_AUTHOR_EMAIL contain "AUTH" (inside "AUTHOR"), so they were being falsely redacted to VAR=*** mangling git commit-authoring commands in tool output (observed when re-authoring commits for the contributor-attribution CI check). Add a leading word-boundary + negative-lookahead allowlist (_GIT_IDENTITY_ALLOWLIST) that exempts the four git identity vars while still redacting every genuine secret. Verified: real keys (OPENAI_API_KEY, AWS_SECRET_ACCESS_KEY, *_AUTH_TOKEN, *_AUTHORIZATION_KEY) still redact; git identity vars (incl. with 'export ' prefix) pass through untouched. Tests: 3 new allowlist tests + 2 still-redact guards; full redact suite 79 passed, 0 regressions. Co-authored-by: Kyzcreig <9063726+Kyzcreig@users.noreply.github.com>

…110 #3) ingest (quarantine) SHALLOW-COPIES rows, so messages[src_idx] is row is ALWAYS False on the normal single-pass path -> the identity guard blocked every stamp. The suffix index is structurally correct (1:1 order-preserving ingest + front-only fold); guard on a cheap role-match instead, and rely on the consumer (in-range validation + reconcile -> A-floor) as the real gate. Test now passes COPIED working rows to prove the stamp fires through ingest copies.

…rtition (#110) * feat(compaction): Option B — provenance-stamped exact in-turn kept partition Makes the in-turn granular announce's folded/kept split EXACT (was A-floor approx, bounded <=7%). Engine stamps _src_idx (origin index into messages) on tail rows before _assemble_context, single-pass only (1:1 mapping holds); the pipeline's shallow-copies carry it through drop/strip/rewrite by construction, stubs lack it. Consumer harvest_provenance_partition() reads exact kept_pre off the returned compressed (NO engine instance state -> no cross-session race, the singleton finding); strip_provenance() removes the key before compressed flows onward (wire/cache guard). Precedence: B (exact) -> replay -> A-floor (approx). Malformed/duplicate/multi-pass/ length-mismatch -> falls to A-floor, never a wrong B split. 13 new tests (provenance partition + engine stamp survival/single-pass-only/strip); 451 compaction + 197 context_engine green. Speculative helpers from #109 removed. * fix(compaction): B provenance — suffix-from-end stamp (Greptile #110 ×2) #1 Provenance never engaged: after a real front-fold, working_messages is shorter than messages even single-pass, so the len-equality guard skipped every production stamp. Fix: the fresh tail is a SUFFIX of the original messages (fold removes from the FRONT only), so index from the END: src_idx = len(messages)-(len(tail)-off), guarded by a per-row identity check (messages[src_idx] is row). New test proves the stamp engages + maps exactly after a 60-row front-fold. #2 Strip-failure could leave provenance in the transcript: made the strip inline (no import that could fail and silently leave _src_idx) — del with pop() fallback. 13 B tests green. * fix(compaction): B stamp guard uses role-match not identity (Greptile #110 #3) ingest (quarantine) SHALLOW-COPIES rows, so messages[src_idx] is row is ALWAYS False on the normal single-pass path -> the identity guard blocked every stamp. The suffix index is structurally correct (1:1 order-preserving ingest + front-only fold); guard on a cheap role-match instead, and rely on the consumer (in-range validation + reconcile -> A-floor) as the real gate. Test now passes COPIED working rows to prove the stamp fires through ingest copies. --------- Co-authored-by: Apollo <apollo@ang.ventures>

Adopts the DNS-rebinding-pinned transport hardening (repo issues #2/#3, PR #3) and the starter-feed/settings failure-surfacing + SSRF-gated icon proxy fix (issue #6, PR #8). Full range in tony-simons-aiowa/hermes-newswire deccdc4..e6b438e (13 commits): Security-relevant highlights: - All outbound fetches (feeds, redirects, icons) now go through a pinned transport: the SSRF gate's validated address set is bound to the actual connection — no second DNS lookup, so DNS rebinding/TOCTOU has no window; the plugin fails closed if the pin seam changes. - New GET /icon.json proxies favicons through the same gate and returns base64 data URLs — the renderer's <img> no longer performs unpinned DNS resolutions of feed-controlled hostnames. 64 KB cap enforced mid-transfer; image content-type allowlist; bounded, normalized TTL cache. - Renderer surfaces backend failures (settings/sources banners, starter-feed inline errors) instead of silent no-ops. Capabilities unchanged (all empty — dashboard plugin, no tools/hooks/ env). Verification at the new pin: 129 pytest, 33 renderer interaction checks, 26 ESM render smoke, hermes plugins validate clean.

FleetReview round 3 on PR #821 raised three P1s against 8cead1e. P1 #1 "watchdog starves teardown" — REFUTED BY MEASUREMENT, with a regression test so it cannot become true silently. The premise is that the inner leash is "drain plus a small grace", so a 35s cap would arm a ~40s watchdog and leave persistence ~5s. DEFAULT_SHUTDOWN_WATCHDOG_GRACE_S is 60.0, not small, so under launchd the inner term never wins the min() and the armed deadline is clamp - LAUNCHD_HARD_EXIT_RESERVE_S. Measured at the production clamp (60): drain 35 -> armed 50 (window 15, need 15); drain 28 with a 22s record -> armed 50 (window 22, need 22). The P1's own proposed assertion passes by construction, both sides being clamp - 10. The arming arithmetic is extracted to resolve_armed_shutdown_watchdog_delay so the invariant is measured against the expression gateway.run really arms with, at both arming sites (live arm + diagnostic snapshot). test_stop_arms_the_watchdog_at_the_hard_exit_deadline drives the real stop() and reads the wall-clock delay handed to arm_shutdown_watchdog — not a call count, not re-derived arithmetic. The grace branch does bind on a clamp far above the drain (clamp > drain + 70); clamp=300 is pinned too, so shrinking the grace to the "small grace" the P1 assumed turns the production row red instead of flipping it quietly. P1 #3 "unbounded teardown reserve" — FIXED at the read/record boundary, not with a drain floor (Argus's floor withdrawal stands). Two bounds: * record_teardown_timing(budgeted=...) marks a stop that actually ran under a supervisor deadline. An unconstrained stop (hermes gateway stop, Ctrl+C, foreground) can legitimately take far longer than any launchd budget; reading that back as the reserve zeroes the next drain. Unbudgeted samples are still written for diagnostics but never read back. Legacy records with no provenance field are not trusted. * read_last_teardown_seconds(max_seconds=...) rejects a sample larger than resolve_max_actionable_teardown_reserve_s(clamp) — the window that exists before the hard exit. At clamp 60 a 55s record drove the drain to 0.0 and stayed poisoned if that stop hard-exited before recording a new sample; it now degrades to "no measurement" (drain 35.0) and is replaced by the next real measurement. P1 #2 "cron drain eats teardown" is DUPLICATE-OF t_f753b2b5 / PR #835 and is not fixed here; the stale `# 45` comment in the cron-leash test is corrected to `# 35` and cites the card. Also replaces test_stop_path_arms_via_the_shared_resolver, which read GatewayRunner.stop's source with inspect.getsource — banned outright by AGENTS.md ("Never read source code in tests") — with the behavioral stop()-driving test above. Verified: - Mutation-checked, each applied to source and reverted: arm from resolve_shutdown_watchdog_delay(drain, grace_s=5.0) -> armed 33.0 vs 50.0, 1 failed; drop the budgeted guard -> 2 failed; drop the max_seconds ceiling -> 2 failed. - Invariant sweep over clamp {None,1,4,8,10,20,30,45,50,60,90,120,300,600} x configured {5,20,30,50,180,600} x last_teardown {None,0,5,22,45,58, 1e9,inf} = 672 combos through the production boot read: 0 violations of window >= max(15, measured) wherever the drain is non-zero; the 240 residuals are all the specified nonnegative saturation (drain == 0); armed < clamp everywhere; non-launchd passthrough intact. - Focused: 46 + 18 pass. Broad shutdown/restart surface, 19 files: 383 passed, 0 failed, 4 skipped. - ruff clean on all changed files; git diff --check clean. Not verified and not claimed: the live safe-restart at host load >= 20 with 5 active turns ending in a clean exit status, the gateway-exit-diag teardown line on a real shutdown, merge-queue landing, deployment, the external fleet-config-lint headroom assertion, and the upstream PR.

… path binds session, every status writer guarded - no caller session identity (plain shell, cron opener) => guard skipped, one stderr note; behaviour identical to base (Argus #1) - _caller_session_id: explicit slash session, then ContextVar-first resolve_current_session_id, env only outside the gateway; gateway /kanban passes the invoking session_id into run_slash (Argus #2) - guard schedule, reopen, reopen-review, request-review, request-changes, link (child), specify, decompose; _home_session_guarded binds the card param by name via inspect.signature; tool handlers for request_review, request_changes, link bound + foreign_ok (Argus #3) - AST contract tests: every tasks.status/assignee/priority/session_id writer is guarded or listed in EXECUTION_LANE with a reason; every CLI verb and tool handler reaching a guarded writer binds the actor Verified: test_kanban_home_session.py 56 passed; 7/8 mutations killed (8th equivalent: CLI + db sessionless checks are redundant layers); kanban suite 146 files: 2074 passed, 3 skipped, 1 failed (workspace_retention[dir_card_states2], passes isolated on head and base).

…ndex for the whole rebuild (t_d3963974 follow-up to #966) Third defect of freeze #3, not covered by #966: repair_external_content_fts ran DROP/CREATE VIRTUAL TABLE/'rebuild'/DROP+CREATE TRIGGER as separately autocommitted statements (legacy isolation opens no implicit txn for DDL). For the 13-28 min rebuild other connections saw messages_fts missing ('no such table', live 05:10) or empty (docsize=0 vs 2.79M on disk). Now: BEGIN IMMEDIATE .. one COMMIT, rollback on error; no-op loads still take no write lock. test_lcm_fts_atomic_rebuild.py: RED on dc228d5 (engine-load structural path torn; failed rebuild left index dropped). The explicit-repair path passed only because #966's parity-marker write opens a txn first. tests/context_engine 397 passed.

… path binds session, every status writer guarded - no caller session identity (plain shell, cron opener) => guard skipped, one stderr note; behaviour identical to base (Argus #1) - _caller_session_id: explicit slash session, then ContextVar-first resolve_current_session_id, env only outside the gateway; gateway /kanban passes the invoking session_id into run_slash (Argus #2) - guard schedule, reopen, reopen-review, request-review, request-changes, link (child), specify, decompose; _home_session_guarded binds the card param by name via inspect.signature; tool handlers for request_review, request_changes, link bound + foreign_ok (Argus #3) - AST contract tests: every tasks.status/assignee/priority/session_id writer is guarded or listed in EXECUTION_LANE with a reason; every CLI verb and tool handler reaching a guarded writer binds the actor Verified: test_kanban_home_session.py 56 passed; 7/8 mutations killed (8th equivalent: CLI + db sessionless checks are redundant layers); kanban suite 146 files: 2074 passed, 3 skipped, 1 failed (workspace_retention[dir_card_states2], passes isolated on head and base).

…sed replays reported lost (t_43e058b7) (#961) * fix: retain session prompt across route metadata writes (#958) Verified targeted session-state, restore, accounting, and model-resume tests: 236 passed, 1 pre-existing dashboard-auth fixture warning deselected. Reproduced NULL from billing route before fix. Co-authored-by: Kyzcreig <9063726+Kyzcreig@users.noreply.github.com> * fix(gateway): restart follow-ups keep adapter-granted admission; refused replays reported lost (t_43e058b7) Argus r8 N1 (t_e253d9d5): SessionSource.to_dict drops is_bot / role_authorized / delivered_via_upstream_relay / profile_route_rejected, so a spooled follow-up admitted only by ALLOW_BOTS, ALLOWED_ROLES or the relay was refused as "Unauthorized user" on boot replay, its spool file acked, and restart_followup_lost logged 0 lines. Trust model: to_dict stays wire-safe (unchanged). The spool record carries the flags in a separate `admission` block and the whole record is HMAC-SHA256'd with a per-home 0600 key (<home>/gateway/restart_followups.key). On load the flags are restored only if the MAC verifies; otherwise no trust flag is restored (only fail-closed profile_route_rejected is honoured) and PHASE=restart_followup_untrusted is logged. Live policy is still re-evaluated by the normal intake. A replay the intake refuses (unauthorized / profile_route_rejected) now logs PHASE=restart_followup_lost with reason. MF (same review): AST contract that the post-turn draining site spools pending_event itself, not None. Verified: new real stop->boot e2e (human/bot/role/relay, forged, tampered, gate-closed-during-restart, to_dict class guard) 8/8; on base 3 admission arms fail, human control passes. Focused restart suites 49/49. Mutants: MAC unchecked, refusal unreported, admission unrestored, MF pending_event=None all KILLED. Argus probe_r8_source_authz_real_intake: B/R PRESERVED, CONTROL ok. Session/authz/startup-restore suites 445 passed. * fix(kanban): judge goal deliverables before completion; fail open for operator errors (#960) * fix(kanban): grade goal deliverables before completion and isolate judge errors Verified 81 targeted tests pass (one ACP-dependent test excluded). Mutating the completion rubric makes the first-completion regression fail as expected. * refactor(kanban): one shared goal-mode handoff gate for CLI and tool surfaces Argus r1 (t_c4e23682): _goal_mode_handoff_rejection was byte-identical in tools/kanban_tools.py and hermes_cli/kanban.py; only the tool copy was test-gated, so 4 CLI mutants survived (Issue NousResearch#38367 two-copies class). - goals.kanban_handoff_rejection is now the single predicate (judge with completion_handoff=True; owned worker retries then blocks transient on judge error; operator fails open with a judge_error event; caller's conn). - Both surfaces' complete + request-review delegate to it, injecting only their run-id resolver and judge-availability probe. - CLI tests drive the real `kanban complete` / `request-review` argv path (build_parser -> kanban_command): completion_handoff reaches the judge and the card closes; real judge prompt accepts first completion; owned-worker 500 -> 2 calls, blocked transient, error on stderr, rc!=0; review gated. - AST contract: exactly one function in the tree calls the judge with completion_handoff, and both surface wrappers delegate to it. Verified: 86 passed, 1 deselected (inherited ModuleNotFoundError: acp, also red on ce0c9d3). Mutation matrix: baseline green precondition, 16/16 KILLED by named failing tests (M01-M13 re-targeted at the shared helper + W1-W9 wiring/duplicate-predicate mutants). --------- Co-authored-by: Kyzcreig <9063726+Kyzcreig@users.noreply.github.com> * fix(kanban): resume dependency-wait PR and page stranded ready cards (#952) * fix(kanban): respawn guard honors worker dependency_wait->promoted resume; add requeue_task + stuck-guard probe t_7d7ff489. Rule 4 active_pr no longer strands a card whose own dependency block (kind=dependency) postdates the newest PR comment and whose promotion has not yet spawned. requeue_task emits operator-intent 'requeued' for READY cards. respawn_guard_stuck_tasks lists cards held by active_pr >= 30 min. * fix(kanban): surface guarded ready cards and provide requeue verb Verify dependency_wait promotion dispatches once and subsequent crash is guarded; CLI requeue and one-shot alert tests pass (168 passed, 1 skipped in focused suites). t_7d7ff489. * test(kanban): keep corruption probe independent of watcher call count Verified: corrupt-board regression 2 passed, 22 deselected; ruff and diff check pass. Original two failures reproduced on clean fork base. * fix(kanban): keep PR continuation through status comments and unobserved ticks Verified 213 passed, 2 skipped across focused DB/CLI/watcher suites; subprocess stdin guard passed. * fix(kanban): consume event-ordered PR requeue intent Verified 183 passed, 1 skipped across DB/CLI/watcher; ruff and subprocess stdin guard pass. Same-second event-order mutant fails the regression test. * docs(kanban): state one-shot PR intent ordering * fix(kanban): bind PR comments to events and consume dependency intent Verified real dispatcher regressions RED before fix, then 210 passed, 2 skipped in focused DB/CLI/watcher suite; ruff and subprocess guard passed. * fix(kanban): pin READY requeue to PR comment identity Legacy same-second inline audit comments can share author and length; requeue snapshots the PR row id and remains one-shot. Verified 211 passed, 2 skipped; ruff clean. * fix(kanban): snapshot comment identity for every resume intent Verified focused DB/CLI/watcher/core suite: 217 passed, 2 skipped. Legacy equal-second strict mutant fails the intended arm. * fix(kanban): guard-stuck age ignores data events; board-scoped recovery command - respawn_guard_stuck_tasks: only kinds that can change the active_pr answer (_RESPAWN_GUARD_FAILURE_RESET_KINDS + dependency_wait + spawned, or a guard decline for another reason) restart the continuous-guard age; comments, heartbeats, attachments are data. - render_operator_command(board, verb, *args): single renderer, always emits --board <slug>; clear_verb uses it; watcher passes the probed board. - test_triage_resolve_records_who_and_why: expect after_comment_id == max(task_comments.id) at emit time (CI slice 15/16 red). Verified: new tests RED on 581299d (4 failed), GREEN here; recovery command executed via real CLI on default and secondary boards (rc0); no-board renderer mutant fails; 244 passed/1 skipped on db/triage/ watchers/cli/boards; ruff clean. --------- Co-authored-by: Kyzcreig <9063726+Kyzcreig@users.noreply.github.com> * fix(lcm): FTS parity COUNT(*) ran under _LOAD_LOCK on every engine load (freeze #3) (#966) * fix(lcm): FTS parity COUNT(*) ran under _LOAD_LOCK on every engine load (freeze #3) Card t_d3963974. Third Apollo boot-cost freeze. The cause: _fts_needs_rebuild_structural ran `SELECT COUNT(*) FROM messages` (SCAN messages USING COVERING INDEX, 2.5M rows) plus `COUNT(*) FROM messages_fts_docsize` on EVERY MessageStore/SummaryDAG construction, under plugins.context_engine._LOAD_LOCK. Measured on an APFS clone of the fleet DB: 13.4 s of a 17.9 s cold MessageStore() init. The two autocommit COUNTs could also straddle a concurrent ingest. They gave a false mismatch 15/300 times, and each mismatch triggered a full inline FTS drop+rebuild. That happened twice on 2026-09-24: held 779.8 s, 12 turns queued. Provenance: the count came in with the original vendor import 8b86963 (2026-06-16) and was carried unchanged through re-vendor 27b6178. #887/#902/#903 did not touch it. #903's plan check exempted USING COVERING INDEX and skipped messages_fts*, so it passed on this code. Fix: - The parity check moves to _fts_count_parity_mismatch. It never runs on the throttle=True (load) path. A metadata marker (fts_parity_checked_at:<fts>) throttles it to once per LCM_FTS_PARITY_CHECK_INTERVAL_HOURS (default 6 h). When due, it runs on the existing background integrity thread (own connection). A mismatch sets the /lcm doctor integrity flag instead of rebuilding inline. Both counts are read in one snapshot. - Explicit repair (throttle=False) and /lcm doctor still run parity synchronously. - A no-op load writes nothing: - _clear_integrity_failed runs only after a real repair. Before, it also erased the background corruption flag on every load. - The messages_dedup_v1 and schema_version upserts are marker-gated. - The integrity claim uses a 1 s busy timeout, not 30 s. - Lifecycle GC (on_session_start, every agent init): - no longer holds BEGIN IMMEDIATE across two SELECT DISTINCT session_id full scans; - uses indexed per-session probes; - runs at most once per 6 h per process. - _backfill_search_content no longer rewrites NULL over NULL for undecryptable rows. That rewrite fired msg_fts_update on every boot. Test: test_lcm_init_cost_regression now traces engine construction + on_session_start on the loading thread. It fails on ANY SCAN of messages, messages_fts*, summary_nodes and nodes_fts*, covering index included (only LIMIT-bounded statements are exempt). It also asserts that a steady-state load needs no write lock and preserves the corruption flag. RED on df43599 (3 failed); GREEN with this change; tests/context_engine: 393 passed. * test(lcm): lock parity-race repro and document restart meltdown --------- Co-authored-by: Apollo <apollo@angventures.io> * fix(gateway): restart follow-ups keep adapter-granted admission; refused replays reported lost (t_43e058b7) Argus r8 N1 (t_e253d9d5): SessionSource.to_dict drops is_bot / role_authorized / delivered_via_upstream_relay / profile_route_rejected, so a spooled follow-up admitted only by ALLOW_BOTS, ALLOWED_ROLES or the relay was refused as "Unauthorized user" on boot replay, its spool file acked, and restart_followup_lost logged 0 lines. Trust model: to_dict stays wire-safe (unchanged). The spool record carries the flags in a separate `admission` block and the whole record is HMAC-SHA256'd with a per-home 0600 key (<home>/gateway/restart_followups.key). On load the flags are restored only if the MAC verifies; otherwise no trust flag is restored (only fail-closed profile_route_rejected is honoured) and PHASE=restart_followup_untrusted is logged. Live policy is still re-evaluated by the normal intake. A replay the intake refuses (unauthorized / profile_route_rejected) now logs PHASE=restart_followup_lost with reason. MF (same review): AST contract that the post-turn draining site spools pending_event itself, not None. Verified: new real stop->boot e2e (human/bot/role/relay, forged, tampered, gate-closed-during-restart, to_dict class guard) 8/8; on base 3 admission arms fail, human control passes. Focused restart suites 49/49. Mutants: MAC unchecked, refusal unreported, admission unrestored, MF pending_event=None all KILLED. Argus probe_r8_source_authz_real_intake: B/R PRESERVED, CONTROL ok. Session/authz/startup-restore suites 445 passed. * fix(gateway): reject torn spool keys and gate replay refusals Verified: 57 focused restart tests passed; forged invalid-key, tampered-admission, and two refusal-site arms exercised via real restart. --------- Co-authored-by: Kyzcreig <9063726+Kyzcreig@users.noreply.github.com> Co-authored-by: Apollo <apollo@angventures.io>

Kyzcreig merged commit 5e6854b into main Jun 3, 2026
19 of 20 checks passed

Kyzcreig deleted the fix/redact-git-identity-allowlist branch June 3, 2026 01:52

Kyzcreig mentioned this pull request Jul 26, 2026

docs(sync): land 2026-07-23 parity evidence (forkdelta review + 257 acks + gate records) #439

Merged

Kyzcreig mentioned this pull request Aug 10, 2026

test(fd-drain): fix the FLAKY burn_fds fixture that made the high-fd guard red #561

Merged

Kyzcreig mentioned this pull request Sep 20, 2026

ci: per-slice runs-on — self-hosted first, GitHub-hosted overflow (one wave instead of ~2.5) #766

Merged

Kyzcreig restored the fix/redact-git-identity-allowlist branch September 21, 2026 10:32

Kyzcreig mentioned this pull request Sep 23, 2026

fix(lcm): ingested_at backfill = second full-table scan on every boot; gate one-time backfills behind lcm_migration_state + lint the init path #902

Merged

Kyzcreig mentioned this pull request Sep 24, 2026

fix(lcm): FTS parity COUNT(*) ran under _LOAD_LOCK on every engine load (freeze #3) #966

Merged

Kyzcreig mentioned this pull request Sep 24, 2026

fix(lcm): make FTS repair one transaction — DDL autocommit tore the index for the whole rebuild (t_d3963974, follow-up to #966) #971

Closed

This was referenced Sep 26, 2026

fix(kanban): auth-time pin refusal uses the card pin, not any --provider #1060

Merged

fix(ci): slice-4 ejections — kernel mock-proc memory leak + O(n^2) launchctl pattern (t_b303235e) #1183

Merged

Navigation Menu

fix(redact): exempt git identity vars from ENV-assignment redaction - #3

fix(redact): exempt git identity vars from ENV-assignment redaction#3
Kyzcreig merged 1 commit into
mainfrom
fix/redact-git-identity-allowlist

Kyzcreig commented Jun 3, 2026

Uh oh!

github-actions Bot commented Jun 3, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

Kyzcreig commented Jun 3, 2026

Problem

Fix

Verification (empirical, not just reasoning)

Risk

Uh oh!

github-actions Bot commented Jun 3, 2026

🔎 Lint report: fix/redact-git-identity-allowlist vs origin/main

ruff

ty (type checker)

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

🔎 Lint report: `fix/redact-git-identity-allowlist` vs `origin/main`