The Agent Skills Directory

[SAFE]: The skill is a set of natural language instructions guiding the AI through a specific investigation methodology. It does not contain executable code, scripts, or external dependencies.
[COMMAND_EXECUTION]: The skill utilizes a restricted set of read-only tools (Read, Glob, Grep). There are no instructions or capabilities present that would allow for arbitrary command execution, privilege escalation, or modification of the underlying system.
[DATA_EXFILTRATION]: The skill does not contain any network-related operations, hardcoded credentials, or external URL references. Data processing is confined to the local environment and the tools specified.
[PROMPT_INJECTION]: The instructions are designed to manage the AI's internal reasoning process (assumption management and "Analytical Overlay") rather than attempting to bypass safety filters or override system instructions.
[INDIRECT_PROMPT_INJECTION]: While the skill is intended to process untrusted data from unknown codebases (ingestion point: Read/Grep operations), the risk is mitigated by the read-only nature of the allowed tools and the protocol's emphasis on separating raw observations from analytical conclusions. No specific capability to execute or write data exists, limiting the impact of any potential injection in the analyzed files.

remote-viewing