Annonce d’un SkillCoordonner des tâches entre agents
Résumé de la publication · anglais
surgical-change scopes Codex edits to a few files when implementing a small feature or bounded bug fix, one of 16 engineering skills in the Boundiff toolkit.
Check user-visible behavior against acceptance criteria using an available local app, staging system, API, or CLI. Prefer read-only smoke tests; never mutate production or external data without authorization.
Critique an existing technical design or proposed architecture for complexity, interfaces, failure modes, and migration cost. Read-only design review; not an implementation request.
Reproduce a reported bug, establish meaningful regression coverage when feasible, apply a minimal fix, and verify it. Use for debugging and regression work; not for general features or read-only audit.
Safely upgrade a specified package or small dependency group, inspect supply-chain changes, and verify compatibility. Use for targeted version updates; not automatic bulk maintenance.
Update user or maintainer documentation to match observed software behavior and commands. Use for documentation drift and feature handoffs; avoid speculative or unrelated rewriting.
Turn a vague feature idea into a small, testable scope and explicit non-goals before code changes. Use for product discovery and unclear requests; not for a fully specified small fix.
Produce an executable, repository-grounded implementation plan with bounded steps and verification gates. Use for multi-file features or risky changes before coding; not for trivial edits.
Perform an independent, read-only review of a commit, branch, pull request, or code path and report reproducible findings. Use for review/audit requests; never silently remediate the target.
Plan or implement a backward-compatible schema or data migration with rollout, validation, and recovery gates. Use for persistent-state changes; never run against production without explicit authorization.
Diagnose a measured latency, memory, throughput, or resource regression and propose the smallest supported optimization. Use for performance questions; measure before tuning.
Prepare a concise, evidence-backed pull request description and reviewer handoff from a known diff. Use after implementation; do not create or publish a PR unless explicitly requested.
Inspect or create a high-integrity Git commit/release checkpoint with exact repository provenance and staged-diff verification. Use for checkpoint, commit, push-readiness, or release preparation; writes require explicit authorization.
Investigate intermittent bugs, regressions, errors, or failing CI to isolate a defensible root cause before changing code. Diagnosis-only; use bug-fix-verification for remediation.
Perform a bounded, read-only application security review of code, dependencies, or a change set. Use for threat and control analysis; not penetration testing of external targets or silent remediation.
Implement a small feature or bounded bug fix with minimal files and a focused diff. Use for scoped implementation; not for broad redesigns, read-only reviews, or release operations.
Review existing test coverage for important behavior, missing assertions, and false confidence without changing code. Use to assess a test suite; not to implement new tests.
Langue du texte source : anglais
mis à jour
Affichage de 16 skills collectés sur 16.
Installer avec un assistant IA
Copiez ce prompt dans l’assistant IA que vous utilisez.
Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/itsofficialadi/boundiff.
