The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes content from an external repository which may contain malicious instructions.
Ingestion points: The read_file and read_frontmatter commands in scripts/gh_kb_helper.sh retrieve content from a GitHub repository into the agent context.
Boundary markers: The skill lacks explicit boundary markers or system instructions to ignore instructions embedded within the retrieved files.
Capability inventory: The helper script includes create and create-from-file commands that allow the agent to write back to the repository, providing a potential path for an injection to persist or spread.
Sanitization: No sanitization or validation of the content retrieved from the repository is performed.
[EXTERNAL_DOWNLOADS]: The skill interacts with the GitHub API via the gh CLI tool to browse and retrieve repository content. These operations are performed against a well-known service (GitHub) and are consistent with the skill's stated purpose.

git-kb-retrieve