Audits workspace files against configured folder, naming, and document rules.
npx skills add https://github.com/baek-labs/hames --skill index-baek-labs
Please help me install this Agent Skill. Skill: index Source: https://github.com/baek-labs/hames/tree/main/packages/claude/skills/index Command: npx skills add https://github.com/baek-labs/hames --skill index-baek-labs
runtime/index.js audit against the configured AI work root, checking placement, naming, and semantic document rules across all workspaces or a single target. - Authorized index repair: Generates a repair plan with runtime/index.js plan, presents the exact proposed changes with a plan_hash, and applies them only after explicit user approval via runtime/index.js apply --approved --plan-hash. - Safe failure handling: Preserves manual text and links, reports broken references instead of rewriting them, and never deletes original files when indexing fails. - Use Case: After weeks of ad-hoc file creation, invoke /index to see which documents violate folder rules, which need human judgment, and which indexes are stale—then approve a bounded repair plan. ## Quick Start Ask the AI to run /index and audit whether all workspace files follow their configured folder and naming rules.High-intent search queries and answers about installing and using this skill.
Invoke /index or ask for a file organization audit. The skill resolves the configured AI work root and runs runtime/index.js audit, optionally scoped with --target to one workspace, then reports violations, judgment items, and unreadable files separately.
Request an index repair explicitly. The skill runs runtime/index.js plan, shows the exact proposed index changes with a plan_hash, and only after your approval runs apply with --approved --plan-hash. It repairs index blocks only and re-audits afterward.
No. The audit is strictly read-only and never repairs indexes, moves documents, renames files, or deletes anything. Even an authorized repair only updates index blocks; moving or editing original documents requires a separate specific request.
The skill requires a configured AI work root. If the root is unconfigured, it directs you to run /setup first rather than inventing folders or rules, since audits depend on the registered workspace configuration.
Semantic document rules require reading actual content, and a path check alone cannot prove meaning. The skill flags uncertain, unreadable, or unsupported items separately and asks you to confirm judgment calls rather than pretending compliance.
automagik-devAutomate initial Genie workspace setup with interactive user prompts.
AgentSystemLabsReorganize ungrouped files into grouped layouts while updating references.
Jose-Luis-NunezDiagnose agent enforcement mechanisms and report FIRED, NOT FIRED, or NOT APPLICABLE per mechanism.
kinetasCreate Harness Engineering workspace folders and baseline operational files.