The Agent Skills Directory

[COMMAND_EXECUTION]: The repository analysis phase in references/generation-process.md provides a series of shell commands for the agent to execute, including ls, grep, tree, and wc. These are used to introspect the technology stack, directory structure, and configuration of the codebase.
[INDIRECT_PROMPT_INJECTION]: The skill represents an indirect injection surface because it ingests untrusted data from the local repository (file paths and content via grep) and incorporates this information into the generated documentation.
Ingestion points: Local repository files and directory structures accessed via shell commands.
Boundary markers: The templates provide structural sections but lack explicit warnings to the agent to ignore instructions embedded within the codebase being analyzed.
Capability inventory: The skill utilizes shell command execution (bash) for analysis and file-write capabilities to generate the AGENTS.md files.
Sanitization: There is no evidence of sanitization for the content extracted from the codebase before it is interpolated into the generated documentation templates.

hierarchical-agents