gtm-motions
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill operates entirely within the agent's defined context, using logical steps to process business inputs. No evidence of obfuscation, remote code execution, or unauthorized data exfiltration was found.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local markdown files, which represents an indirect prompt injection surface. However, given the skill lacks network or command execution capabilities, the risk is negligible.
- Ingestion points: Reads sections 2, 3, and 4 from
/foundation/brain.mdand guardrails from/context/meta-patterns.mdin Step 0. - Boundary markers: The skill does not use explicit delimiters or "ignore instructions" warnings when loading context files.
- Capability inventory: The skill generates text and tables, and performs restricted writes to
/context/skill-sessions.md. It has no access to shell commands, network operations, or dynamic code evaluation. - Sanitization: No specific sanitization of external content is implemented before interpolation into the prompt.- [DATA_EXPOSURE]: The skill reads business-related context files (
/foundation/brain.md). This access is within the intended scope of the skill for GTM analysis and does not involve sensitive system paths, private keys, or credentials.
Audit Metadata