bluesky-dev-playbook
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No attempts to override system prompts or bypass safety filters were detected. The skill defines its own scope and limitations clearly, instructing the agent to refuse requests related to coding bots or integration engineering.
- [DATA_EXFILTRATION]: There are no patterns involving reading sensitive files or sending local data to external servers. The skill focuses on public social media strategy and uses standard web-based references.
- [OBFUSCATION]: The content consists of clear, human-readable Markdown. No Base64, homoglyphs, zero-width characters, or hidden text patterns were found.
- [REMOTE_CODE_EXECUTION]: The skill does not perform any remote code execution. It explicitly directs users to official documentation for technical integrations, remaining strictly within the realm of social media policy and strategy.
- [DYNAMIC_CONTEXT_INJECTION]: No use of dynamic shell command execution (e.g., backtick syntax) was found in the documentation.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user queries about social media strategy. While this is an ingestion point for external data, the risk is negligible because the skill does not have any capabilities to execute code, write files, or perform network operations. It also includes boundary markers such as specific labeling requirements ([Bluesky], [research]) for its outputs.
Audit Metadata