The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands including ls, find, and grep during its 'Step 0: Project Status Perception' phase to scan the local wiki/ directory. These commands are used to identify the presence of documentation and the status of various project items.
[COMMAND_EXECUTION]: The skill instructions specify that the agent should use a local script, read-yaml.mjs, located in a sub-skill directory (cc-init/references/tools/), to extract fields from YAML files rather than reading them directly.
[PROMPT_INJECTION]: The skill identifies and processes data from the wiki/ directory to guide its routing logic. Maliciously crafted metadata or status fields within these project files could influence the agent's workflow transitions.
Ingestion points: The wiki/ directory and its subdirectories (e.g., requirements/, features/, roadmaps/) are scanned for status markers and stale flags.
Boundary markers: No explicit delimiters or instruction-bypass warnings are defined in the routing logic.
Capability inventory: Uses ls, find, and grep to inspect files and routes execution to other sub-skills (e.g., cc-spec, cc-fix) based on the findings.
Sanitization: The skill does not define sanitization or validation routines for the content read from the wiki/ files before using it for routing decisions.

cc