openpress-collaborate

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes content from project source files to generate and execute edit proposals.
  • Ingestion points: Project source files (e.g., .mdx and .tsx files in the press/ directory) and @openpress-comment markers are read to inform proposals.
  • Boundary markers: The skill does not explicitly define delimiters to separate ingested file content from the agent's instructions.
  • Capability inventory: The agent can write to the handoff file (.openpress/review/current.json) and perform source file edits via the 'Apply' and 'Direct' modes.
  • Sanitization: The 'Apply' mode requires explicit user confirmation and performs exact string matching of 'before' text before applying any changes, ensuring the user has oversight of all modifications.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 11:18 AM
Security Audit — agent-trust-hub — openpress-collaborate