feat: Phases 8-9 + catalog pipeline refactor#831
Merged
Conversation
…lict The Component interface in registry.ts (crawled catalog entries) conflicted with the main Component interface in component/types.ts. Renamed to CatalogRecord with a distinguishing doc comment. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds toEntityType() and toComponentType() bridge functions to convert between CLI kebab-case ComponentType and KG snake_case EntityType, enabling consistent type mapping across the codebase. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
readConfigFile() now validates parsed TOML against a valibot schema before transforming keys. Invalid configs warn but still work (lenient mode). Exports AgentsConfigTomlSchema for future `agents config validate`. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract Plugin class, validation functions, and display helpers from commands/plugin.ts into lib/plugin-ops.ts for reuse from lint.ts. Wire up plugin linting in lint.ts via dynamic import of plugin-ops. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- ai-txa: Rename registry.ts Component → CatalogRecord to disambiguate - ai-567: Add toEntityType() / toComponentType() bridge functions - ai-a4s: Add valibot schema validation to TOML config reader - ai-78a: Extract Plugin class to lib/plugin-ops.ts (plugin.ts 1354→658), wire lint.ts for plugin validation 9 new tests, zero regressions Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
… subcommands Implements CLI configuration management on top of the existing config.ts foundation. Supports dotted-path key access, auto type coercion, layered source attribution, and $EDITOR integration. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add generic initComponent() that copies template dirs with {{name}}
replacement. Wire into `agents init` for types with templateDir.
Templates: content/personas/_template, content/hooks/_template.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ling Replace components/skills -> content/skills, context/* -> content/*, cli/ -> packages/cli/ in CLAUDE.md, README.md, content files, justfiles, and Python embed scripts. Update CLAUDE.md directory structure to reflect current layout (packages/cli/src/, hooks/, package.json, bun.lock). Update CLI grammar references from noun-first to verb-first. Fix ruff.toml exclude/ignore paths for content/ directory rename. Add missing words to cspell dictionary. Fix pre-existing markdown lint. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Remove unnecessary escape sequences in bash/zsh completion generators. The \$ escapes before non-brace characters are redundant in JS template literals and flagged by biome's noUselessEscapeInString rule. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…fixes - ai-zgs: `agents config get/set/list/edit/path` with TOML storage - ai-9ev: Scaffold templates for persona and hook types + generic initComponent() for template-based scaffolding - ai-m44: `agents completions bash/zsh/fish` shell completion generator - ai-6d6: Fix 42 files with stale path references (components/→content/, context/→content/, cli/→packages/cli/src/), update CLAUDE.md + README 28 new tests, zero regressions Closes: ai-zgs, ai-9ev, ai-m44, ai-6d6, ai-kh8 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Checks runtime (Bun, Node, Git), project structure, config validity, and installed component counts. Supports --json output mode. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Serve supports --web (graph viewer, default) and --api (minimal REST endpoints: /api/health, /api/components, /api/components/:type). Wires both doctor and serve into the agents CLI entrypoint. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- `agents doctor`: Runtime, project, config, and component health checks with structured output (human/JSON) - `agents serve --web`: Delegates to existing graph-viewer - `agents serve --api`: Minimal Bun HTTP server with /api/health, /api/components, /api/components/:type endpoints 22 new tests, zero regressions Closes: ai-62i Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…wnloadBatch Replace per-skill downloadBatch() with per-repo discoverAllRepos() in the analyze subcommand. Repos with multiple skills are now cloned once during a single discovery phase, then looked up from a map during batch processing. - Add content field to DiscoveredSkillResult so SKILL.md text survives cleanup - Store content in computeAllMechanicalFields() during discovery - Analyze now runs discovery once upfront, builds source:skill lookup maps - processBatch resolves skills from discovery map instead of cloning - Propagate all mechanical fields from discovery (including ComponentMetadata) - Remove unused downloadBatch and cpSync imports from catalog.ts Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
… (ai-1f1) The discover subcommand now runs reconcile() + detectMoveRenames() on discovery results and prints a reconciliation summary. When --apply is passed, applyReconciliation() writes changes back to the catalog NDJSON: adds new entries, marks removals, refreshes mechanical fields on updated/moved entries, and handles renames preserving analysis data. Writes are atomic (tmp + rename). Also adds --full flag to force re-cloning all repos (bypass cache). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add clarifying comments to downloadSkill(), downloadBatch(), and backfillEntries() in catalog-download.ts noting they serve the `agents add skill` use case, not catalog analysis (which now uses discoverAllRepos from catalog-discover.ts). Notes backfillEntries as a future candidate for discovery-based refactoring. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Refactor analyze to use discoverAllRepos() instead of downloadBatch(), cloning each unique repo exactly once instead of per-skill - Store SKILL.md content in DiscoveredSkillResult for manifest building - Add applyReconciliation() + --apply flag to discover subcommand - Add --full flag to bypass incremental discovery cache - Document download vs discovery code path separation - Incremental HEAD SHA checking was already wired (pre-existing) Zero test regressions, same NDJSON output format Closes: ai-hns, ai-1f1, ai-n1k, ai-o8m, ai-f3u Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
agents configcommand (get/set/list/edit/path), scaffold templates for persona/hook types, shell completions (bash/zsh/fish), fix 42 stale path references across docsagents doctor(environment health checks) +agents serve(graph-viewer web + API endpoints)discoverAllRepos()as primary entry point;--applyflag for auto-reconciliationTest plan
agents config set general.debug true+agents config get general.debuground-tripsagents completions bashoutputs valid completion scriptagents doctorshows runtime/project/config/component checksagents catalog discover --applyapplies reconciliation🤖 Generated with Claude Code