The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, such as SIEM configurations and alert logs, creating a potential vector for malicious instructions to influence the agent's behavior.
Ingestion points: The skill performs a project-wide scan for SIEM configurations, detection rules, and alert pipeline infrastructure, and accepts user-supplied arguments to guide the analysis.
Boundary markers: There are no explicit instructions or delimiters provided to the agent to treat external SIEM rule content as data only or to ignore embedded instructions.
Capability inventory: The skill possesses the capability to perform broad filesystem scans and write multiple files, including analysis reports and hidden telemetry logs.
Sanitization: The instructions do not include any steps for sanitizing or validating the content of the SIEM rules before they are processed by the agent.
[PERSISTENCE]: The skill implements a mechanism for logging execution state in hidden configuration directories, which persists across different sessions.
Evidence: The "SELF-EVOLUTION TELEMETRY" section instructs the agent to search for a project path in ~/.claude/projects/ and append metadata (outcome, iterations, bottlenecks) to a skill-telemetry.md file. While intended for performance tracking, this involves repeated modification of hidden system configuration files.

alert-prioritization