Fix residual gaps: typed hooks, security hardening, CLI testability#8
Merged
ApiliumDevTeam merged 26 commits intodevfrom Mar 3, 2026
Merged
Fix residual gaps: typed hooks, security hardening, CLI testability#8ApiliumDevTeam merged 26 commits intodevfrom
ApiliumDevTeam merged 26 commits intodevfrom
Conversation
## Summary - **Bump pnpm overrides**: `minimatch` 10.2.1→10.2.3 (2x HIGH ReDoS), `fast-xml-parser` 5.3.6→5.3.8 (LOW stack overflow) - **Bump node-llama-cpp** 3.15.1→3.17.1: drops 4 deprecated transitive packages (npmlog, gauge, are-we-there-yet, old tar) - **Drop @discordjs/opus**: eliminates 7 deprecated packages from the `node-pre-gyp` chain; `opusscript` is already the preferred codec - **Fix 107 README links**: maryos→mayros across all docs - **Fix 15 broken internal doc links** caught by `docs:check-links` - **Fix 15 test files for vitest 4.x** mock hoisting compatibility (BlueBubbles, Telegram, Slack, Feishu, agent-runner, cron, hooks/install, CortexSidecar) - **Track CI-required assets** (.gitignore negation rules for template docs, a2ui bundle placeholder, test fixture tgz) - **Bump version to 0.1.2** + sync 38 extensions - Platform updates: iOS, Android (Play Store + sideload flavors), macOS - Rename clawnet→meshnet in refactor docs ## Test plan - [x] `pnpm install` — no resolution errors - [x] `pnpm build` — succeeds - [x] All 239 previously-failing tests pass locally (12 test files verified) - [x] CI: check-docs, build-artifacts, CodeQL, skills-python, protocol, android test/build all green - [ ] CI: full test suite on push - [ ] CI: macOS job 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Bumps the docker-images group with 2 updates: node and debian. Updates `node` from 22-bookworm to 25-bookworm Updates `debian` from `98f4b71` to `74d56e3` --- updated-dependencies: - dependency-name: node dependency-version: 25-bookworm dependency-type: direct:production dependency-group: docker-images - dependency-name: debian dependency-version: bookworm-slim dependency-type: direct:production dependency-group: docker-images ... Signed-off-by: dependabot[bot] <support@github.com>
Bumps the docker-images group with 2 updates: node and debian. Updates `node` from 22-bookworm to 25-bookworm Updates `debian` from `98f4b71` to `74d56e3` Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting `@dependabot rebase`. [//]: # (dependabot-automerge-start) [//]: # (dependabot-automerge-end) --- <details> <summary>Dependabot commands and options</summary> <br /> You can trigger Dependabot actions by commenting on this PR: - `@dependabot rebase` will rebase this PR - `@dependabot recreate` will recreate this PR, overwriting any edits that have been made to it - `@dependabot show <dependency name> ignore conditions` will show all of the ignore conditions of the specified dependency - `@dependabot ignore <dependency name> major version` will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself) - `@dependabot ignore <dependency name> minor version` will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself) - `@dependabot ignore <dependency name>` will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself) - `@dependabot unignore <dependency name>` will remove all of the ignore conditions of the specified dependency - `@dependabot unignore <dependency name> <ignore condition>` will remove the ignore condition of the specified dependency and ignore conditions </details>
The deriveIdHint() function was returning "mayros-bluebubbles" from the npm package name, but plugin manifests use short IDs like "bluebubbles". Strip the "mayros-" prefix so hints match manifests and eliminate 38 config warnings on startup. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
…t timeout - Android bundle-playstore job now skips gracefully when KEYSTORE_FILE secret is not configured, instead of failing with ENOENT - Increase Discord pairing test timeout from 10s to 60s to handle slow CI runners under heavy load Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
## Summary - **Strip `mayros-` prefix from plugin entry hints** — eliminates 38 config warnings on startup by aligning derived hints with manifest short IDs - **Skip Android `bundle-playstore` when keystore is missing** — job now skips gracefully instead of failing with ENOENT - **Increase Discord pairing test timeout** — 10s→60s to handle slow CI runners ## Test plan - [x] Plugin discovery tests pass (8/8) - [x] Discord monitor tests pass (3/3) - [x] All 239 previously-fixed tests still pass - [ ] CI green on push 🤖 Generated with [Claude Code](https://claude.com/claude-code)
## Summary - Bump root package.json version to 0.1.3 - Sync 38 extension versions to 0.1.3 ## Test plan - [x] `pnpm plugins:sync` — 38 updated, 0 skipped - [ ] CI green 🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- llm-task, open-prose: use mayros/plugin-sdk instead of relative path - llm-task, open-prose: add missing devDependency on @apilium/mayros - googlechat, memory-core: fix peerDependency version to >=0.1.0 - skill-hub: add missing semver dependency Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Makes CLI exit points testable by routing through the runtime abstraction instead of calling process.exit directly. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- command-registry: pass argv param instead of using process.argv - register.subclis: wrap eager registration in Promise.allSettled to prevent a single failing subcli from breaking others Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Move ALLOWED_INVALID_COMMANDS and ALLOWED_INVALID_GATEWAY_SUBCOMMANDS to shared lightweight-commands module - Add shouldRegisterPrimarySubcommand guard in run-main - Add resetConfigGuardForTest helper Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use pickGatewayPort() instead of hardcoded 18789 so the SSH tunnel command matches the actual configured gateway port. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Allows tests to reset the bannerEmitted state between test cases. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- clearAgentRunContext: also clear seqByRun entry to prevent stale sequence numbers across runs - resetAgentRunContextForTest: clear all Maps (runContextById, seqByRun, listeners) for complete test isolation - Add missing resetAgentRunContextForTest() call in 3rd test Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use { hash?: string } instead of any for the createTriple no-op stub.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add 3 new patterns: 'do not follow system/developer', 'developer message', and XML-tag injection (<system>, <assistant>, etc.) - Add 7 tests for uncovered regex variants: forget, pretend, you should, run bash, override instructions, wget, eval Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- prepareContext: batch listTriples calls in groups of 5 with Promise.all to reduce sequential Cortex round-trips - mergeResults: pre-fetch all parent text values into a Set to eliminate N+1 dedup queries - Add removeInjectedContext to prevent memory leaks from accumulated delegation contexts Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Store companion triple with full original predicate alongside conflict-flag triple for unambiguous resolution - resolveConflicts: look up companion triple for exact original predicate, with endsWith fallback for legacy data - Clean up companion triples when resolving conflicts Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Using owners[0] as grantor was incorrect — the first owner cannot grant admin to themselves. Use 'system' as the grantor instead. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Hooks: - subagent_spawning: use typed event.agentId/label/childSessionKey instead of as Record casts - subagent_ended: use event.targetSessionKey/outcome/runId, clean up injected context on end - before_agent_start: mark unused event as _event - agent_end: use typed event.success instead of cast Tools: - mesh_share: enforce namespace prefix on subject and predicate - mesh_query: two-step query via ownership triples instead of single patternQuery that could leak cross-namespace data - mesh_conflicts: remove always-true conditional after early return - CLI share: add write access check before createTriple Infra: - Bound messageLog to 1000 entries via appendToMessageLog helper - Add ensureNsPrefix utility for consistent namespace enforcement Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Remove dead before_agent_start hook (event.skills does not exist on PluginHookBeforeAgentStartEvent) - Remove before_tool_call hook and toolCallTimers map; rewrite after_tool_call using typed event.toolName/params/result/durationMs - Fix LLM hooks: use event.runId instead of callId, event.usage.input and event.usage.output instead of promptTokens/completionTokens - Remove redundant String() wrapper on event.model (already string) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- semantic-skills before_tool_call: use typed event.toolName directly - semantic-skills after_tool_call: use typed event.toolName and event.result directly, remove redundant null guard - memory-semantic before_compaction: use typed event.messages directly instead of casting to Record<string, unknown> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The test exercises /bash via handleCommands but did not call resetBashChatCommandForTests() first, risking stale activeJob state from prior tests. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
ApiliumDevTeam
added a commit
that referenced
this pull request
Mar 9, 2026
## Summary - **Rewrite extension hooks** to use typed event fields instead of `as Record<string, unknown>` casts across semantic-observability, semantic-skills, memory-semantic, and agent-mesh (eliminates 13 unsafe casts) - **Harden agent-mesh**: namespace-prefix enforcement on share/query tools, bounded message log, ACL grant fix, companion triples for conflict resolution, batched delegation queries - **Replace `process.exit`** with `defaultRuntime.exit` in 4 CLI modules (hooks, logs, plugins, help) for testability - **Fix CLI command registration**: pass argv explicitly, use `Promise.allSettled` for subcli registration, extract lightweight-commands constants - **Add injection detection patterns** (3 new) and **variant test coverage** (7 new tests) in enrichment-sanitizer - **Fix agent-events state cleanup**: clear seqByRun + listeners in reset, add missing test resets in agent-events and commands tests - **Fix extension packages**: import paths (`mayros/plugin-sdk`), missing devDependencies, peerDependency versions 17 atomic commits, 33 files, 406 insertions, 260 deletions. ## Test plan - [x] `npx tsc --noEmit` — 0 errors - [x] `pnpm build` — clean - [x] `pnpm test -- src/infra/agent-events.test.ts extensions/semantic-skills/enrichment-sanitizer.test.ts src/auto-reply/reply/commands.test.ts` — all pass - [x] `grep -n "as Record" extensions/semantic-observability/index.ts` → 0 matches - [x] `grep -n "(event as Record" extensions/memory-semantic/index.ts` → 0 matches - [x] `grep -n "process\.exit" src/cli/hooks-cli.ts src/cli/plugins-cli.ts src/cli/logs-cli.ts src/cli/program/help.ts` → 0 matches - [x] Full test suite `pnpm test` — no regressions
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
as Record<string, unknown>casts across semantic-observability, semantic-skills, memory-semantic, and agent-mesh (eliminates 13 unsafe casts)process.exitwithdefaultRuntime.exitin 4 CLI modules (hooks, logs, plugins, help) for testabilityPromise.allSettledfor subcli registration, extract lightweight-commands constantsmayros/plugin-sdk), missing devDependencies, peerDependency versions17 atomic commits, 33 files, 406 insertions, 260 deletions.
Test plan
npx tsc --noEmit— 0 errorspnpm build— cleanpnpm test -- src/infra/agent-events.test.ts extensions/semantic-skills/enrichment-sanitizer.test.ts src/auto-reply/reply/commands.test.ts— all passgrep -n "as Record" extensions/semantic-observability/index.ts→ 0 matchesgrep -n "(event as Record" extensions/memory-semantic/index.ts→ 0 matchesgrep -n "process\.exit" src/cli/hooks-cli.ts src/cli/plugins-cli.ts src/cli/logs-cli.ts src/cli/program/help.ts→ 0 matchespnpm test— no regressions