The Agent Skills Directory

[PROMPT_INJECTION]: The skill manages a potential indirect prompt injection surface where untrusted repository data is processed.
Ingestion points: Repository files and overlay commentary (SKILL.md).
Boundary markers: Present; instructions specify that overlay commentary, explain docs, and proposed cross-links are advisory and should not be treated as canonical source truth.
Capability inventory: Search, read, and potential source mutation via synrepo_apply_anchor_edits across various skill tools.
Sanitization: Relies on the external synrepo tool logic and explicit agent instructions to prioritize authoritative graph facts over advisory content.
[COMMAND_EXECUTION]: The skill uses local synrepo CLI commands for repository analysis. All mentioned commands and flags (such as --allow-source-edits) are standard for the tool's documented purpose.
[DATA_EXFILTRATION]: No exfiltration patterns or network requests to non-whitelisted domains were detected; operations are restricted to the local environment.

synrepo