internal-comms
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFENO_CODEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructions create a surface for indirect prompt injection by directing the agent to ingest and summarize untrusted data from external platforms. \n
- Ingestion points: The documentation in
examples/3p-updates.md,examples/company-newsletter.md, andexamples/faq-answers.mdinstructs the agent to gather content from Slack, Email, Google Drive, and Calendar events. \n - Boundary markers: There are no boundary markers or specific instructions provided to the agent to treat external content as untrusted or to ignore embedded instructions. \n
- Capability inventory: The agent uses its core summarization capabilities to process the data into newsletters and status reports. No external tool execution or file system writes are defined within the skill files themselves. \n
- Sanitization: No sanitization, escaping, or validation steps are prescribed for the retrieved content before it is interpolated into the generated communication. \n- [NO_CODE]: This skill is composed solely of Markdown documentation and a license file, with no executable code, scripts, or package manifests.
Audit Metadata