The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses several shell and git commands to perform discovery and record metadata.
Evidence includes instructions to run git status, git rev-parse, git diff, git log, git show, and rg --files to identify source changes.
It also uses cat to write metadata to openwiki/.last-update.json.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it processes untrusted data from the repository to influence agent behavior through instruction files.
Ingestion points: The skill reads repository tree structures, source files, and git commit history from the local environment (SKILL.md, core.md).
Boundary markers: While it uses a specific markdown template for the OpenWiki reference section in AGENTS.md and CLAUDE.md, it does not specify delimiters for other processed content.
Capability inventory: The agent has the ability to read arbitrary repository files, execute shell commands, and write to both the openwiki/ directory and root-level instruction files like AGENTS.md and CLAUDE.md.
Sanitization: No explicit sanitization or filtering of content extracted from source files is mentioned before it is written into the documentation or root instruction files.

wiki-update