xlfg-engineering

Audit xlfg workflow load, SDLC coverage, Claude Code compatibility, and benchmark readiness.

Autonomous xlfg diagnosis run. Batches hidden recall, intent, context, and debug skills to find the deep root problem without changing source code.

Manual bootstrap / repair for xlfg scaffolding in this repo.

Autonomous xlfg SDLC run. Batches hidden recall, intent, context, plan, implement, verify, review, and compound skills end-to-end.

Adversarial task critic. Use proactively after each implementation task to catch drift before phase completion. Owns one atomic lane and returns only after the required artifact is complete.

Scoped patch engineer. Use proactively for each non-trivial task so implementation follows the planned root fix. Owns one atomic lane and returns only after the required artifact is complete.

Regression-proof builder. Use proactively whenever tests or proofs must change to match a task. Owns one atomic lane and returns only after the required artifact is complete.

Solution-space explorer for ambiguous requests. Use proactively when /xlfg needs concrete options before committing to a plan. Owns one atomic lane and returns only after the required artifact is complete.

Adjacent-requirement hunter. Use proactively before planning to surface implied behaviors and parity gaps. Owns one atomic lane and returns only after the required artifact is complete.

Constraint auditor. Use proactively before planning to surface runtime, dependency, security, and ops constraints. Owns one atomic lane and returns only after the required artifact is complete.

Ambiguity triager. Use proactively before planning to isolate assumptions, deferrable decisions, and the few true blockers. Owns one atomic lane and returns only after the required artifact is complete.

Verification harness medic. Use proactively when local servers, ports, health checks, or flaky dev environments matter. Owns one atomic lane and returns only after the required artifact is complete.

Budget and proof profiler. Use proactively to choose the lightest honest verification harness for a run. Owns one atomic lane and returns only after the required artifact is complete.

Intent surgeon for messy requests. Use proactively at the start of /xlfg to turn vague asks into a concrete intent contract. Owns one atomic lane and returns only after the required artifact is complete.

Repository cartographer. Use proactively at the start of /xlfg to map structure, commands, and conventions. Owns one atomic lane and returns only after the required artifact is complete.

External fact-check researcher. Use proactively when current docs, APIs, or domain facts could change the plan. Owns one atomic lane and returns only after the required artifact is complete.

Risk and rollback auditor. Use proactively for risky changes to expose release, safety, and proof pressure points. Owns one atomic lane and returns only after the required artifact is complete.

Causal debugger. Use proactively during planning to find the true failure mechanism before solutioning. Owns one atomic lane and returns only after the required artifact is complete.

Root-fix architect. Use proactively after diagnosis to compare options and pick the smallest honest fix. Owns one atomic lane and returns only after the required artifact is complete.

Behavior-contract author. Use proactively when flows need concrete scenario-level behavior before coding. Owns one atomic lane and returns only after the required artifact is complete.

Atomic task packet planner. Use proactively after solution choice so each specialist gets one bounded mission, one artifact, and one honest done check. Returns only after the artifact is complete.

Proof-gate reviewer. Use proactively before implementation to decide whether the test contract is READY or REVISE. Owns one atomic lane and returns only after the required artifact is complete.

Proof-contract engineer. Use proactively during planning to define the minimum honest fast checks and ship checks. Owns one atomic lane and returns only after the required artifact is complete.

UI design and a11y contract specialist. Use proactively in planning and verify lanes when the task touches frontend or visual/interaction design. Owns one atomic lane; returns only after the artifact is complete.

User-value anchor. Use proactively during planning to keep the run aligned to real operator impact and false-success traps. Owns one atomic lane and returns only after the required artifact is complete.

Architecture critic. Use proactively for maintainability, layering, and contract-fidelity review. Owns one atomic lane and returns only after the required artifact is complete.

Performance critic. Use proactively for hot paths, latency, and slow-harness traps. Owns one atomic lane and returns only after the required artifact is complete.

Security critic. Use proactively for auth, validation, secret-handling, and injection review. Owns one atomic lane and returns only after the required artifact is complete.

UX and accessibility critic. Use proactively for scenario fidelity, usability, and a11y review. Owns one atomic lane and returns only after the required artifact is complete.

Verification evidence judge. Use proactively after verify-runner to reduce logs into decisive run truth. Owns one atomic lane and returns only after the required artifact is complete.

Verification operator. Use proactively to execute layered checks and capture real evidence artifacts. Owns one atomic lane and returns only after the required artifact is complete.

Internal xlfg phase skill. Use only during /xlfg runs to promote verified durable lessons into knowledge without copying the entire run into memory.

Internal xlfg phase skill. Use only during /xlfg runs to gather repo truth, current constraints, harness facts, and targeted external research when needed.

Internal xlfg phase skill. Use only during /xlfg-debug runs to reproduce the failure, separate symptom from mechanism, and write an evidence-backed diagnosis without changing source code.

Use docs/xlfg and .xlfg as the file-based context system for lean, autonomous SDLC runs.

Internal xlfg phase skill. Use only during /xlfg runs to implement the plan, update tests, and keep the run card truthful without asking the user to code.

Internal xlfg phase skill. Use only during /xlfg runs to resolve messy user intent into a compact contract inside spec.md before broad repo fan-out.

Internal xlfg phase skill. Use only during /xlfg runs to turn intent plus context into a lean run card, atomic task packets, a practical proof contract, and a READY gate.

Apply production-readiness gates to an xlfg run without recreating duplicated planning state.

Internal xlfg phase skill. Use only during /xlfg runs to perform deterministic recall and record the smallest relevant carry-forward memory.

Deterministic recall for xlfg memory. Use current-state first, then temporal or typed lexical recall over durable knowledge, role memory, the ledger, and runs.

Internal xlfg phase skill. Use only during /xlfg runs to run proportional multi-lens review and capture only net-new findings or residual risk.

Internal xlfg phase skill. Use only during /xlfg runs to run layered proof, reduce results into verification.md, and surface the first actionable failure.

3 hooks across 3 events