Agent System Coach
Teach developers how to work effectively with the repository's multi-agent workflow, including mandatory verification, review commands, and controlled rule-refresh cycles.
api-contract-review
Review API/schema/interface changes for backward compatibility and consumer safety.
task-progress
Interactively produce a high-quality English progress comment for tracker tasks, with ArkUI/OpenHarmony risk coverage and optional mapping to structured task descriptions (expected outcome, scope, acceptance criteria).
task-specifier
Transforms raw user ideas into structured, technical task records for tasks.jsonl.
build-system-analysis
Analyze complex multi-tool build systems and produce actionable build topology and change plans.
ios-code-review
Review iOS code for lifecycle safety, concurrency correctness, security, performance, and test gaps.
android-code-review
Review Android code for lifecycle safety, threading, correctness, security, performance, and test gaps.
java-code-review
Review Java code for correctness, concurrency safety, security, performance, and test completeness.
<skill-name>
<one-line skill summary>
ci-failure-debugger
Diagnose CI failures and produce minimal, verifiable remediation plans.
OpenHarmony Task Specifier
Help developers write one high-quality OpenHarmony task description by asking guiding questions and producing a clear final draft with expected outcome, scope, verification, and domain-specific risks.
large-codebase-context
Keep agent context small and evidence-grounded while navigating and changing very large codebases.
code-review
Review code changes via local diff, branch diff, or latest commit with severity-ranked findings and concrete fixes.
security-review
Review changes for exploitable vulnerabilities, trust-boundary issues, and insecure secret handling.
performance-review
Review changes for latency, throughput, memory, allocation, and scalability regressions.