rss-ai-digest

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill operates by executing a local command-line tool named subscription-research or digestpilot-runtime. It contains instructions for the agent to resolve this command using environment variables (DIGESTPILOT_RUNTIME_CMD) or by executing repository-specific bootstrap scripts such as scripts/doctor.mjs and scripts/bootstrap-runtime.mjs.
  • [EXTERNAL_DOWNLOADS]: The skill's primary function is to download and process content from external RSS and Atom feeds. It includes a curated OPML file containing various technical and AI-related feed URLs from established sources.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is designed to ingest and summarize untrusted data from the internet.
  • Ingestion points: External RSS/Atom feeds and OPML files defined in references/feed-registry.md and references/base-feeds.opml.
  • Boundary markers: Not present; there are no specific instructions for the agent to use delimiters or ignore embedded instructions within the feed content.
  • Capability inventory: Execution of the subscription-research CLI and file system operations for managing registries and state files.
  • Sanitization: Not specified; the skill relies on the agent's internal summarization logic to process the fetched data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 11:53 PM