The Agent Skills Directory

[SAFE]: The skill is primarily instructional and focused on establishing a secure qualitative research methodology. It does not contain any executable code, scripts, or commands that perform network operations, file system modifications outside of a designated workspace, or privilege escalation.- [DATA_EXPOSURE]: The instructions include explicit safeguards for handling sensitive project data. It mandates the use of an 'isolated workspace' for private observations and strictly forbids writing credentials, source identifiers, or original script fragments into the public skill tree or promotion ledgers.- [INDIRECT_PROMPT_INJECTION]: While the skill is designed to ingest external text sources (scripts and project chains), it incorporates a comprehensive defense-in-depth strategy. This includes 'de-copying' (abstracting content to remove unique identifiers), blind forward evaluation by a 'fresh agent' to prevent instruction leakage, and final approval by an independent reviewer.- [PROMPT_INJECTION]: No evidence of prompt injection or safety bypass patterns was found. The skill reinforces safety by instructing the agent to stop reading if authorization cannot be verified and to prioritize synthetic materials for practice.- [OBFUSCATION]: The skill uses clear, technical language throughout. There is no evidence of Base64 encoding, hidden Unicode characters, or other obfuscation techniques designed to mask malicious intent.

short-drama-knowhow