agently-collaboration

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs an anonymous GET request to the GitHub API (api.github.com/repos/AgentEra/Agently-Skills) to check for metadata updates. This operation is limited to version comparison and specifically forbids the loading or execution of remote skill bodies.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests external project data, such as task descriptions and document links, to create summaries and progress reports.\n
  • Ingestion points: External task titles, document URLs, and requirement descriptions provided by the user or fetched from the project environment as defined in references/taskboard-and-handoffs.md.\n
  • Boundary markers: The skill relies on structured formatting like Markdown tables and specific identifiers, though it lacks strict delimiters for untrusted external content.\n
  • Capability inventory: Reporting tools and the specific network tool for versioning; the skill excludes automated file writing or command execution capabilities during standard operation.\n
  • Sanitization: Focuses on human-readable translation and accuracy rather than technical sanitization of input strings.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 05:57 AM