The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute git log commands for metadata retrieval, specifically ranking files by recency. This usage follows best practices (using the -- separator) and is restricted to repository discovery.
[INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and analyzing untrusted source code, creating a surface for indirect injection.
Ingestion points: Local source files are accessed via the Read, Glob, and Grep tools (documented in SKILL.md).
Capability inventory: The agent can read files, search directories, query language servers, and execute reasoning chains.
Boundary markers: The skill mitigates risks through a structured 6-phase protocol and mandatory file:line citations, though it does not explicitly instruct the agent to use XML delimiters or ignore instructions within the analyzed code.
Sanitization: No explicit sanitization or filtering of the source code content is performed during the reading phases.

deep-read