The library

judge-panel · Assembling diverse expert perspectives in a structured panel verdict

deepcheck · Deep verification of code correctness with multi-lens checking

design-review · Systematically reviewing design artifacts across consistency and usability

path-warden-completeness · Verifying task completion paths are complete and well-formed

tdd-cycle · Testing TDD workflows when you need automated red/green/mutation gates without interactive review stops.

athena-pr-reviewer-workflow · Auditing PRs when you need parallel specialist analysis (comment, test, error, type, code, simplify, requirements) with batched verification.

fs-harness-research · Researching filesystem abstractions when comparing Flue, Hare, cloudflare-agents, Mastra, and Pi implementations.

codebase-audit · Auditing codebases when you need whole-repo cross-slice analysis (orphans, duplication, dead code, architecture drift, HARD-RULE compliance).

hiraia-expand-wave · Building fact banks when you need medium-grained concept beats expanded into distinct trilingual (Tagalog/English/Bisaya) child-grade facts.

inspector-usability-audit · Auditing UI form builders when you need full data-flow verification (form→state→serialization) against upstream schema docs.

content-rendering-qa · QA text content when you need to detect HTML/entities, page numbers, OCR gibberish, dropped words, truncation in rendered fields.

liftoff-workflow · Project execution when you need parallel Astronauts to implement task batches with FC verification and worktree commits.

«family»-agent-headtohead · Benchmarking agent variants when you need head-to-head evaluation on a fixed task set.

lang-content-generate · Content localization when translating a spec into multiple languages (often with cultural/regional variants).

refactor · Improving code quality when tests are passing and you want to safely restructure without behavior change.

optimization-loop · Improving performance when you need goal-driven iteration (profile → hypothesis → implement → measure).

wf-plan · Designing workflows when you need to decompose a goal into agent-ready phase briefs and schemas.

deep-insilico-program-design · Algorithm/architecture design when you need iterative sketches stress-tested by adversaries before coding.

skill-improver-batch · Skill refinement when you have many skills needing updates based on failure logs or user feedback.

evaluate-external-workflow · Validating imported workflows when you need to verify they work on your task set.

feature · Feature delivery when you need end-to-end orchestration (design → implement → test → deploy).

test-analysis · Debugging tests when you need to distinguish flakiness from real failures and identify patterns.

trigger-eval-harden · Automating testing when you need to harden Workflow triggers before production deployment.

unit-test-gen · Test coverage when you have code without tests and need to bootstrap test suite.

load-test · Finding capacity limits through stepped load with automatic breaking-point detection.

fixture-dynamic-roster-missing-skill · Multi-phase orchestration with agent parallelism and structured verification.

toolkit-improvement · Multi-phase orchestration with agent parallelism and structured verification.

wf-brainstorm · Synthesizing design decisions across conflicting expert perspectives.

fierce-ralph · Iterative task execution with real-time progress verification and stagnation detection.

cheap-pass-100 · High-volume batch screening with fixed concurrent worker capacity.

trace-resume-model · Distinguishing prefix vs. content-addressed workflow resumption models.

feature-dev · Multi-gate feature development with human approval at requirements and architecture phases.

finish-l1-primitives · Refactoring monolithic code into bit-identical primitives with gated extraction.

parallel-file-write-with-verify · Writing many files in parallel then auditing for quality metrics.

gate · Code review with parallel specialists and adversarial skepticism.

spec-critique-voice-ladder-regen · Criticizing specifications from multiple evaluation angles with voice consistency.

go-audit-fix · Automated auditing and fixing of Go code issues.

type-field-coverage · Measuring type coverage across large codebases.

sdd-implement-engine · Implementing complex systems driven by formal specifications.

triage_quarantine · Isolating failing cases for targeted root-cause investigation.

docking-workflow · Integrating components with automatic verification.

fd-understand-codebase · Building knowledge of unfamiliar codebases through structured exploration.

assess-investigate · Code review with parallel specialists and adversarial skepticism.

phase1-repos · Processing multiple repositories in a phased workflow.

g3-readiness-review · Rigorously vetting academic manuscripts through multiple independent reviewer lenses before submission, with explicit synthesis of conflicts and prioritized fixes.

fathomdb-agent-memory-impl-strategy · Grounding implementation strategy for agent-memory gaps in the actual engine code, with adversarial re-verification against real mechanisms and invariants.

polymarket-reverse-engineer · Designing independent, first-principles trading strategies from observed edge patterns, falsifiable through rigorous gauntlet testing, without relying on wallet-copying.

drain-wave-1 · Draining large spec items with sub-agent isolation and orchestrator-side independent verification of runtime results, preventing sub-agent report blindness.

insilico-feasibility-and-track-record · Honestly evaluating which in-silico predictions have real clinical track records versus hype, with explicit real_world_status per example.

phase1e-scan · Surfacing external unknowns that could block Phase 2 launch through execute-access probing (RUN code, not just read), with optional skeptic verification gated on FAIL only.