commercial-proposal-writer
Pass
Audited by Gen Agent Trust Hub on Feb 23, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious instructions, obfuscation, or safety-bypass attempts were identified in the skill documentation or reference files.
- [NO_CODE]: The skill consists entirely of Markdown instructions and structural templates. It does not include any Python, Node.js, or shell scripts, eliminating the attack surface for remote code execution or privilege escalation.
- [DATA_EXFILTRATION]: The skill processes local project files for context but provides no mechanisms or commands (such as curl or wget) to transmit data to external servers.
- [PROMPT_INJECTION]: An indirect prompt injection surface exists as the skill ingests untrusted data from various local markdown files and user inputs. 1. Ingestion points: multiple local files (solution-brief, discovery-notes, etc.) and user_input. 2. Boundary markers: absent. 3. Capability inventory: local file-writing (commercial-proposal, workplan). 4. Sanitization: absent. The risk is minimized by the absence of dangerous system-level capabilities.
Audit Metadata