orgs-ai-harness

Use this skill when adding or fixing orgs-ai-harness tests, validation rules, generated artifact schema checks, or local quality commands for the Python unittest suite.

Use this skill when changing draft approval, protected artifact metadata, eval replay, verification thresholds, repo-local cache refresh, export policy, or explain output.

Use this skill when modifying GitHub repository discovery, `gh repo list` integration, explicit repo selection, fork/archive filters, clone behavior, or interactive checkbox selection.

Use this skill when changing orgs-ai-harness CLI commands, setup wizard flows, argparse subcommands, user-facing command output, or command error handling in `src/orgs_ai_harness/cli.py`.

Use this skill when editing `harness.yml` parsing, repo registry mutations, org pack initialization, repo coverage statuses, or config round-trip behavior in orgs-ai-harness.

Use this skill when adding or fixing orgs-ai-harness tests, validation rules, generated artifact schema checks, or local quality commands for the Python unittest suite.

Use this skill when modifying proposal-first learning, trace-backed improvement proposals, source refresh proposals, proposal apply/reject commands, redaction, or stale-cache handling.

Use this skill when changing repo onboarding scans, sensitive file skipping, generated Agent Skills prompts, LLM staging targets, resolvers, evals, draft pack reports, or `onboard` behavior.

Use this skill when an AI coding agent edits ametel01 repos, especially with dirty worktrees, generated skills, CLI harness code, or multi-file changes. It encodes read-before-edit, scoped diffs, command verification, validation-before-completion, and user-change preservation habits observed across the org.

Use this skill in ametel01 TypeScript repos and packages that use Bun, Biome, Prettier, strict tsconfig, Vitest or Bun test, Next.js, Hono, Apibara, or workspace scripts. It guides installs, lint/format/typecheck/build commands, TypeScript invariants, and known formatting variants.

Use this skill when changing GitHub Actions, Makefile/package verification scripts, CI-required checks, build pipelines, or code that must pass local quality gates across ametel01 repos. It maps repo variants for Bun, Go, Scarb, frontend, indexer, and release workflows.

Use this skill when changing README content, docs directories, API contracts, CLI behavior, developer workflows, deployment notes, changelogs, or user-facing behavior in ametel01 repos. It explains how to keep docs aligned with code and CI without inventing unsupported policies.

Use this skill when entering an ametel01 repo or package and choosing install, build, test, lint, format, release, or CI commands. Detect Bun, Biome, Prettier, Go Makefiles, Scarb/snforge, Python packaging, lockfiles, and GitHub Actions before running commands.

Use this skill when preparing releases, version bumps, changelog entries, GitHub tag workflows, GoReleaser changes, generated release notes, or release-gate summaries in ametel01 repos. It detects manual changelog versus tag-driven release variants.

Use this skill when changing dependencies, install/update scripts, secret handling, GitHub Actions security, vulnerability scanning, checksums, provenance, Dependabot, TruffleHog, govulncheck, or frontend script-safety rules in ametel01 repos. Guidance is conditional where evidence is repo-specific.

Use this skill when adding, fixing, or reviewing tests in ametel01 repos using Bun test, Vitest, Playwright, Go tests, Python unittest/pytest-style harness tests, or Cairo snforge tests. It captures shared expectations for regression coverage, deterministic fixtures, and CI-aligned test commands.

Use this skill when changing repo onboarding scans, sensitive file skipping, generated Agent Skills prompts, LLM staging targets, resolvers, evals, draft pack reports, or `onboard` behavior.

Use this skill when changing agent-vitals Commander CLI commands, source option behavior, terminal report output, Markdown report output, compare/health/changelog-facing command docs, or command validation. Keywords: src/index.ts, report, health, compare, annotate, impact, changes, TerminalReport, MarkdownReport.

Use this skill when editing the agent-vitals single-file dashboard, dashboard HTTP API, docs site, README metric explanations, source selectors, date filters, chart data, or user-facing dashboard documentation. Keywords: dashboard.html, serveDashboard, React CDN, Recharts, /api/metrics, docs/index.html.

Use this skill when adding or changing agent-vitals quality metrics, daily metric SQL, rolling regression thresholds, provider-scoped benchmarks, health status, compare output, or metric labels. Keywords: MetricsAnalyzer, daily_metrics, RegressionDetector, read_edit_ratio, blind_edit_rate, Claude-only metrics, provider _all.

Use this skill when changing agent-vitals package metadata, Bun/Node build workflow, install or uninstall scripts, generated Claude slash commands, auto-scan script, published files, or release/version notes. Keywords: package.json, bun.lock, dist/index.js, scripts/install.sh, scripts/uninstall.sh, scripts/auto-scan.sh, CHANGELOG.

Use this skill when changing agent-vitals baseline recommendations, prescribe diagnostics, Claude settings writers, Codex config/AGENTS.md writers, known-fix thresholds, legacy .rules cleanup, or impact records. Keywords: prescribe, baseline, Prescriber, CODEX_KNOWN_FIXES, KNOWN_FIXES, AGENTS.md, config.toml, settings.json.

Use this skill when changing agent-vitals session scanning, Claude or Codex log adapters, tool-call classification, JSONL parsing, source filters, SQLite ingestion, or provider metadata. Keywords: scan, ClaudeAdapter, CodexAdapter, ParsedSessionLog, tool_calls, thinking_blocks, read/edit/search/bash classification, --source.

Use this skill when adding or changing agents-toolbelt catalog tools, install/update/uninstall methods, package-manager support, dependency bootstrapping, tool verification commands, tiers, categories, or skill exposure in `internal/catalog` and `internal/pkgmgr`.

Use this skill when changing agents-toolbelt Cobra commands, `cmd/atb/runtime.go`, Bubble Tea pickers, install/update/uninstall/status/catalog command behavior, stdout/stderr rendering, `--yes` non-interactive mode, or TUI selection/search behavior.

Use this skill when changing agents-toolbelt generation of the `cli-tools` Agent Skill for Claude Code and Codex, including skill targets, default paths, generated SKILL.md content, category ordering, verified tool inventory, or `internal/skill` golden tests.

Use this skill when changing Go code, tests, Makefile quality gates, lint configuration, or developer docs in agents-toolbelt. Covers the repo's Go 1.26 toolchain, `make verify`, strict golangci-lint profile, test expectations, and documentation/changelog updates.

Use this skill when changing agents-toolbelt release packaging, GoReleaser config, GitHub Actions release flow, `scripts/install.sh`, self-update downloads, checksum verification, cosign signing, provenance attestation, version injection, or installer security behavior.

Use this skill when changing agents-toolbelt state reconciliation, install receipts, managed-vs-external ownership, install/update/uninstall planning, verification persistence, shell-hook state, or lifecycle summaries in `internal/state`, `internal/discovery`, `internal/plan`, and `cmd/atb/runtime.go`.

Use this skill when editing Horizon Protocol Cairo contracts, snforge tests, WAD math, PT/YT/SY token flows, Router functions, Market AMM logic, Oracle integration, access control, or upgradeable Starknet components in contracts/src and contracts/tests.

Use this skill when working with Horizon Starknet deployment, declaration, address export, class hash comparison, market initialization, liquidity seeding, devnet/fork Docker stacks, Sepolia/mainnet env files, sncast accounts, or upgrade scripts in deploy/.

Use this skill when changing the Horizon Next.js dApp in packages/frontend, including Starknet wallet hooks, typed contract calls, Router/RouterStatic integrations, FSD feature modules, WAD math, React Query cache keys, shadcn UI, or contract ABI codegen.

Use this skill when modifying the Horizon Apibara indexer in packages/indexer, including Starknet event filters, factory-pattern contract discovery, Drizzle schema or migrations, Zod event validation, u256/i256/ByteArray parsing, knownContracts, health metrics, and PostgreSQL views.

Use this skill when choosing validation commands or CI-equivalent checks for horizon-starknet changes across Cairo contracts, Next.js frontend, Apibara indexer, Docker devnet, generated ABIs, formatting, tests, builds, or GitHub Actions workflows.

Use this skill when adding or changing vitals-db analytics query modules, DateRange handling, DTO/Zod schemas in @vitals/core, query tests, or API response contracts for metrics such as sleep, workouts, HRV, zones, load, energy, running dynamics, and daily aggregates.

Use this skill when modifying vitals-db Apple Health XML ingest, including HK identifiers, saxes streaming parser behavior, crop/rebuild, mappers, unit normalization, UTC timestamps, _ingest_seen deduplication, fixtures, and ingest tests.

Use this skill when changing vitals-db composite analytics, performance reports, readiness, training strain, aerobic efficiency, run economy, fatigue flags, load quality, recovery debt, consistency index, or paid-report answer/evidence/action logic.

Use this skill when changing vitals-db DuckDB schema, ordered SQL migrations, database connection code, analytics tables, indexes, _migrations behavior, or migration tests in packages/db.

Use this skill when adding or changing vitals-db Hono API routes, server env handling, CLI serve behavior, query validation, HTTP error shapes, docs/API_CONTRACT.md, or frontend fetch functions that call apps/server endpoints.

Use this skill when modifying the vitals-db Next.js dashboard under apps/web, including App Router pages, server component data fetching, ECharts chart components, format helpers, API client usage, dashboard, sleep, performance, workouts, and Next 16 behavior.