Profile My skills Saved Collections Edit profile Submit a skill

← all publishers

CHENG-LIANG1

@cheng-liang1 source repo

9 published skills

  1. ▌
    Agent Change Gate · cheng-liang1 bundle
    Review and gate AI-generated code changes by reconstructing intent, mapping diff risk, selecting independent verification, checking security and dependency impact, and requiring evidence before merge or release. Use for coding-agent pull requests, autonomous patches, or large AI-generated diffs; do not use for ordinary implementation requests or style-only review.
    0
    installs
  2. ▌
    Context Architect · cheng-liang1 bundle
    Structure long-running, multi-file, or handoff-prone work into a compact evidence-linked context that survives compaction and agent transitions. Use when a task spans multiple phases, sources, decisions, or sessions; do not use for short self-contained requests.
    0
    installs
  3. ▌
    Failure Forensics · cheng-liang1 bundle
    Diagnose reproducible or intermittent failures in tests, builds, CI, dev servers, CLIs, and integrations by turning symptoms into falsifiable hypotheses and small discriminating experiments. Use when the root cause is unknown; do not use for a routine fix whose cause is already established.
    0
    installs
  4. ▌
    Blast Radius Guard · cheng-liang1 bundle
    Preflight consequential agent actions by resolving exact targets, authority, reversibility, retry ambiguity, rollback, and independent verification. Use before destructive, externally visible, bulk, financial, permission, deployment, send, or publish operations; do not use for read-only work or trivial reversible local edits.
    0
    installs
  5. ▌
    Agent Eval Contract · cheng-liang1 bundle
    Design trustworthy, reproducible evaluations for AI agents, skills, prompts, and tool loops by turning quality claims into scenarios, evidence, baselines, validity checks, and ship gates. Use for version comparisons, regression detection, or release decisions; do not use for ordinary deterministic unit tests.
    0
    installs
  6. ▌
    Agent Harness Auditor · cheng-liang1 bundle
    Audit or improve a code repository's AI-agent harness by making project knowledge discoverable, commands reproducible, architectural invariants executable, and feedback loops fast. Use when preparing a codebase for Codex or other coding agents, investigating inconsistent agent changes, or reviewing AGENTS.md and repository guidance; do not use for ordinary code review or a one-off task with no repository-level issue.
    0
    installs
  7. ▌
    Multi Agent Conductor · cheng-liang1 bundle
    Coordinate multi-agent or subagent work by deciding when delegation helps, partitioning dependencies, assigning exclusive ownership, defining handoff artifacts, and verifying one integrated result. Use only when the user or applicable instructions explicitly request delegation, subagents, or parallel agent work; do not use for ordinary single-agent tasks.
    0
    installs
  8. ▌
    Agentic Workflow Contract · cheng-liang1 bundle
    Design safe, repeatable agentic automations for scheduled or event-driven work by defining triggers, permissions, idempotency, budgets, review gates, failure handling, and observable proof. Use when creating or reviewing recurring AI-agent workflows, Codex automations, or GitHub Agentic Workflows; do not use for one-off agent tasks or deterministic jobs better handled by a normal script.
    0
    installs
  9. ▌
    Instruction Provenance Guard · cheng-liang1 bundle
    Protect AI-agent tasks from indirect prompt injection and untrusted tool output by tracing instruction provenance, separating data from authority, constraining source-to-sink flows, and confirming sensitive disclosure or action. Use when reading webpages, issues, emails, documents, logs, MCP results, or third-party skills while tools can access private data or mutate state; do not use for ordinary trusted inputs with no external-content boundary.
    0
    installs