Agent Skills by Zhang Xindong
Showing 21 vetted skills indexed across 1 GitHub repositories.
eval-fix
Iteratively generate and evaluate fixes until feature gaps converge.
archive
Merge feature deltas into full specs and designs with review and version control.
verify
Run unit, integration, and end-to-end tests and store verdicts under .harness/evidence.
skill-discovery
Analyze Git commits and logs to propose new Harness Skills.
cancel-sprint
Detect and terminate active Harness sprint loops by deleting .harness/sprint-loop.md.
baseline
Capture baseline documentation of web products via Playwright MCP exploration.
sdd-init
Scan a repository and generate .harness config, spec, and design files.
evolve
Extract recurring failure patterns from sprint histories and generate invariants.
adversarial-review
Identify hidden design flaws and edge cases in code submissions with evidence-based reports.
harness
Automates software delivery from idea to verification with spec and test gates.
proposal
Guide structured feature specification with acceptance criteria and technical constraints.
evaluate
Compare reference and development products via scripted tests and screenshots.
tdd-align
Generate three-layer tests from approved specs into executable test contracts.
harness-status
Analyze local Harness state to determine the current development phase.
steward
Enforce module ownership and contract boundaries for code changes against AGENT.md.
entropy-clean
Scan latest wave changes to remove debug artifacts and enforce style.
sprint
Orchestrate autonomous sprint loops executing pending development tasks with governance checks.
spec-review
Collect line-by-line comments and track resolutions across spec reviews.
decompose
Decompose feature-driven test suites into atomic tasks with wave-based parallel execution and module stewardship artifacts.
harness-init
Analyze repository files and team norms to generate .harness configuration artifacts.
help
Explain Harness plugin slash commands for Spec-Driven workflows.