The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external files which can be controlled by an attacker in a repository context.
Ingestion points: The skill instructions in SKILL.md and the format reference in references/formats.md detail the loading of data from SARIF results (results.sarif), weAudit annotation files (.vscode/*.weaudit), and binary graph JSON exports (binary_graph.json).
Boundary markers: There are no explicit boundary markers or instructions to the AI agent to ignore or treat content within these files as data rather than instructions.
Capability inventory: The skill leverages the trailmark library to perform various graph operations including augment_sarif, augment_weaudit, and augment_binary, which involve annotating code nodes and creating new subgraphs based on the external input.
Sanitization: The instructions do not specify any explicit sanitization or filtering of the content within the ingested files, relying on the underlying trailmark engine's internal parsing logic.

audit-augmentation