marketing-context
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructions for 'Auto-Draft from Codebase' (Mode 1) direct the agent to ingest content from various project files such as READMEs, landing pages, and configuration files. This creates a surface for indirect prompt injection, where malicious or misleading content within those files could manipulate the resulting marketing context draft.
- Ingestion points: Project codebase files (README, package.json, landing pages) referenced in
SKILL.mdunder 'Mode 1: Auto-Draft from Codebase'. - Boundary markers: Absent; there are no instructions to the agent to treat ingested codebase content as untrusted or to ignore instructions embedded within those files.
- Capability inventory: The skill is limited to reading project files and generating a markdown document (
.agents/marketing-context.md). It lacks capabilities for network access, arbitrary command execution, or sensitive credential access. - Sanitization: Absent; the skill does not specify any filtering or validation of the content read from the codebase before it is used to generate the positioning draft.
Audit Metadata