Skip to content

arxdsilva/agent-office / .claude/skills/add-event-source

add-event-source

Verified

Add a new agent source/provider to agent-office (e.g. LangGraph, OpenAI Agents SDK, OpenTelemetry GenAI spans, Claude Code hooks). Use when asked to support or ingest agents from another framework or data source.

Add an event source

Rules that apply: TESTING.md (TDD, ≥70%), ARCHITECTURE.md, PRIVACY_SECURITY.md.

First ask: can the producer just POST /api/events? If yes, document it in the README instead of adding code.

  1. Red — packages/adapters/src/sources/<name>/*.test.ts: fixtures shaped like the real data with invented content. Test a pure mapper toEvent(raw, ctx): AgentEvent | null:
    • identity (agentId stable and namespaced, e.g. lg:<id>), parentId for sub-agents
    • provider (the zone) and topic (the row) — what would a user call this conversation?
    • status mapping onto idle | thinking | tool | waiting | error | done
    • malformed input → null
  2. Green — packages/adapters/src/sources/<name>/: a pure mapper plus a driving adapter with start() / stop() / a testable tick(). It depends only on core ports (ReportActivity, optionally AgentInbox, Clock, Scheduler); put file/network access behind an adapter-local interface (like TranscriptFiles) so it can be tested in memory. Never throw out of tick().
  3. Config — a setting in shared/src/desktop-api.ts SettingsSchema (additive, with a default, and a restart field in core/src/domain/settings.ts), plus an env override in adapters/src/config.ts and .env.example. Start or stop it in startBackend in desktop/src/main/composition.ts.
  4. E2E — packages/adapters/test/e2e.<name>.test.ts: real source → real core → observe(office) sees the agent in the right zone and row.
  5. Docs — README "Sources" section: what is read, what is ignored, privacy notes.
  6. Privacy check — only status, tool, title, model and ≤80-char snippets leave the source; nothing is logged raw.
  7. make check (including make arch), then /run-office to see the new zone.

Runtime-based providers (a CLI agent-office can also start and manage, such as Codex or an ACP agent): this skill is not enough. Follow the checklist in docs/provider-integrations.md §8 (declaration, profile or runner, probe, optional observation hooks, composition, contract suites).

reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 469kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 321kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 321kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 321kUpdated Claude CodeCodexVerifiedInstall