tech-marketing-framework
Warn
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DYNAMIC_EXECUTION]: The skill features an '/autoresearch' tool that autonomously updates its own skill definitions based on binary evaluations. This constitutes self-modifying behavior where instructions are altered at runtime based on the success of generated outputs.
- [DYNAMIC_EXECUTION]: The skill includes a '/skill-builder' utility designed to generate new skill definitions, representing a runtime script generation capability.
- [INDIRECT_PROMPT_INJECTION]: The framework is designed to ingest and process data from external and user-provided sources, creating a risk for indirect prompt injection attacks.
- Ingestion points: The skill reads from local files in 'docs/inputs/' (including product briefs, testimonials, and competitor intelligence) and fetches content from external competitor URLs via the '/how-they-market' agent.
- Boundary markers: Absent. The instructions do not specify any delimiters or warnings to the agent for ignoring instructions embedded within the ingested data.
- Capability inventory: The skill has the ability to write files to the 'output/' directory, perform network requests for competitive research and image generation, and execute Git operations (commits) through its optimization feature.
- Sanitization: Absent. There is no evidence of validation, filtering, or escaping of data retrieved from external URLs or local templates before it is used to influence agent prompts.
- [EXTERNAL_DOWNLOADS]: The installation section provides instructions to clone the skill from a third-party repository ('github.com/j1ngg/tech-marketing-framework.git') that is not associated with a known trusted organization or the designated vendor.
Audit Metadata