The Agent Skills Directory

[SAFE]: No security issues were identified during the analysis of the skill's instructions, playbooks, or templates. The skill's primary function is to provide structured guidelines for building safe and robust agent-facing interfaces.
[DATA_EXFILTRATION]: No exfiltration patterns were found. The skill includes a dedicated playbook (references/playbooks/sdk-telemetry.md) that explicitly instructs users to implement content-capture toggles and PII redaction to prevent accidental data leakage.
[COMMAND_EXECUTION]: The skill contains a shell script (evals/run-static-checks.sh) used for static analysis and validation. This script is used for development-time consistency checks and does not perform unauthorized or malicious operations during agent runtime.
[PROMPT_INJECTION]: The skill does not contain instructions to bypass safety filters or override agent constraints. The prompts defined in references/subagent-dispatch.md are intended for internal lens-based analysis and follow the skill's stated diagnostic purpose.

agent-dx