workshop-critique
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists exclusively of instructional text for the AI agent and contains no executable scripts, binaries, or configuration files that could affect the host system.
- [SAFE]: The instructions establish a strict 'cone of silence' and a 'read-only' rule that prevents the agent from rewriting or composing prose, serving as a significant behavioral constraint.
- [PROMPT_INJECTION]: Because the skill processes untrusted user-provided chapters and stories, it possesses an indirect prompt injection surface. However, the agent lacks the necessary system capabilities (such as network access or file writing) to execute malicious instructions.
- Ingestion points: User-provided text for workshop critique in SKILL.md.
- Boundary markers: The instructions do not define explicit delimiters to isolate the text being analyzed from the agent's core instructions.
- Capability inventory: The skill is strictly read-only and lacks network, file-write, or shell access.
- Sanitization: No sanitization or filtering of the input text is performed.
Audit Metadata