The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes project source code and diagram files to generate, update, or audit documentation. This creates a surface for indirect prompt injection where malicious instructions embedded in these files could influence the agent's subsequent actions.
Ingestion points: The agent is instructed to use Read and Glob tools to ingest content from project directories such as src/ and ai/diagrams/ as described in SKILL.md and references/AUDIT.md.
Boundary markers: The skill relies on Markdown headers and Mermaid code blocks for structure but does not include explicit delimiters or instructions to the agent to disregard natural language commands found within the source files.
Capability inventory: The agent has access to Bash, Write, and Edit tools, granting it the ability to modify the filesystem or execute shell commands based on its analysis of the ingested data.
Sanitization: No explicit sanitization or input validation mechanisms are defined for the content processed by the skill.

diagram