h4tcher-capability-discover

Capability discovery detects overlapping artifacts by performing a read-only scan of your repository to identify existing capabilities, collision risks, and inbound reference counts before you modify code. It cross-references agents, skills, rules, commands, hooks, and adapters.

A read-only preflight scan for capability changes is a safe repository investigation that verifies a clean working tree and maps existing artifacts without writing anything. It produces a structured report identifying duplication risks and governance pillar gaps.

Mapping governance gaps when adding new agents or hooks involves cross-referencing canonical inventory and repository surfaces to identify which governance pillar a new artifact would close. The discovery process triages scope and emits a bounded-length report with cited file-path evidence.

Capability discovery supports multi-surface investigations across pipelines, adapters, agents, skills, rules, commands, and governance files. It applies preflight triage by change scope to select the right investigation depth and optionally dispatches sub-agents for higher-tier research.

You can run a duplication scan without modifying your working tree because the capability discovery process enforces a strict read-only workflow. It verifies a clean working tree before performing any cross-reference analysis or web research.

You should avoid using an automated capability discovery report when you need to execute actual code changes, because the process is strictly read-only and bounded. It is designed for preflight scoping and collision mapping, not for writing artifacts to the repository.