The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and interpret external, untrusted data in the form of git diffs and patch files. This creates a surface where malicious instructions could be embedded within code changes to influence agent behavior.
Ingestion points: The agent is instructed to read and analyze git diff output and patch files as part of the primary workflow in SKILL.md.
Boundary markers: Absent; there are no explicit instructions for the agent to treat the diff content as data only or to ignore any linguistic instructions found within the hunks.
Capability inventory: The skill is configured with access to Bash, Read, and Grep tools.
Sanitization: No validation, escaping, or filtering is applied to the content of the diffs before they are processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill metadata references a external repository for its skill graph context.
Evidence: The skill_graph_source_repo field in SKILL.md points to https://github.com/jacob-balslev/skill-graph, which is the author's official repository.

diff-analysis