apm
01Agent
Personas are Opus-only. Angelique, APM — adversarial junior-PM plan reviewer. ELI5 the choice, then challenge the explanation. Plans only, never code or results.
5 3d ago A 43 tokens
5Stars on the repository
69Mods indexed here, across every type
3d agoLast push, which is what freshness is scored on
noneNo LICENSE: all rights reserved, so bodies are not copied
Agent
Per-page clarity verdict for one architecture atlas page — followability, boundary-entry specificity, ambiguity risk. Never accuracy/citations (C4's mechanical job); never rewrites the page.
5 3d ago A 44 tokens
Agent
Weekly-gate reviewer, static-only, never executes. Writes chunk- .md incrementally, surviving compaction. Only via parallel-code-review.
5 3d ago A 37 tokens
Agent
Static-only Sonnet reviewer, never executes: nits, weak tests, security. Sidecar findings, returns pointer + OK/WARN/BLOCKED verdict.
5 3d ago A 36 tokens
Agent
Fresh-eyes auditor cross-referencing specialist claims against a finished synthesis; read-only, binary present/absent verdict.
5 3d ago C 29 tokens
Agent
Audits dependency manifests for CVEs (npm/pip/cargo/govulncheck) to disk. Manifests only — security-audit-worker scans source instead.
5 3d ago A 42 tokens
Agent
Validates documentation links — internal file/anchor existence plus external URL checks. Returns a broken/redirected/timeout/ok table.
5 3d ago A 32 tokens
docs-checker
08Agent
Verifies external API claims against authoritative docs (Context7, LSP) before an expensive Opus review. A table, not a review.
5 3d ago A 33 tokens
eng-director
09Agent
Personas are Opus-only. The Director of Engineering, Director of Engineering — the Staff Engineer-level rigor plus cross-team/cross-repo boundary authority. Ask-the-sibling bias.
5 3d ago E 41 tokens
enricher
10Agent
Enriches plan stubs pre-execution; maintains live plan bodies and registers mid-execution. Gathers and verifies facts, never decides.
5 3d ago A 33 tokens
executor
11Agent
Implements enriched, reviewed stub specs — the typist, not the architect. Validates at chunk boundaries, stops on unclear specs.
5 3d ago A 29 tokens
Agent
Authors and runs the baseline falsifier for a plan's prime exit criterion — the instrument, never the spec. Sees the prime exit criterion and the repo only; never the ACs, chunk bodies, or task spine.
5 3d ago C 52 tokens
Agent
Opt-in bounded web research on plan claims prior-art-checker left Silent, on empirically-tricky topics. Sidecar signal buckets; never mutates the plan.
5 3d ago C 38 tokens
Agent
The one dispatchable committer — verifies a supplied pathspec, commits via the sanctioned scoped route. Refuses any unbounded or missing pathspec.
5 3d ago A 35 tokens
Agent
Haiku NotebookLM scout — finds video/podcast/article sources per notebook topic, writes sources.md for workers. Discovery only.
5 3d ago A 33 tokens
Agent
Personas are Opus-only. Waste — Kira: is this code too much? Overengineering, spaghetti, redundant work, structures that survived because they existed. Never correctness.
5 3d ago A 41 tokens
Agent
Synthesizes chunk-reviewer outputs plus 3 specialist workers into a BLOCKED/WARN/OK verdict. Invoked only by parallel-code-review.
5 3d ago C 37 tokens
Agent
Mechanical checker: does a plan's fix slate cover its audit oracle, are deferrals justified and ratified, is the task-spine resolved. Never auto-fixes.
5 3d ago A 40 tokens
Agent
Recall pre-flight cross-referencing a plan or research question against wikis, decisions, and lessons before an Opus review. Sidecar: Conflicts/Compatible/Silent.
5 3d ago B 40 tokens
repo-scout
20Agent
Haiku repo-research scout — inventories assigned repo chunks for Sonnet specialists. Mechanical inventory only.
5 3d ago C 24 tokens
Agent
Sonnet repo-research specialist — deep-reads a scout's file inventory, challenges peer claims, writes verified claims.json.
5 3d ago C 29 tokens
Agent
Haiku web-research scout — runs scope.md's queries, vets accessibility, writes the shared source corpus. Discovery only.
5 3d ago A 29 tokens
Agent
Sonnet web-research specialist — deep-reads a scout's source corpus, verifies claims, writes claims.json + summary.md.
5 3d ago C 30 tokens
Agent
Opus NotebookLM sweep — blocked until workers finish, assesses claim coverage, fills gaps, frames the final document.
5 3d ago C 27 tokens