blogwatcher
Warn
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions include installing a Go binary from a third-party GitHub repository (
github.com/Hyaxia/blogwatcher/cmd/blogwatcher). This source is not verified as a well-known service or official organization. - [COMMAND_EXECUTION]: The skill performs its primary functions by executing the
blogwatcherCLI tool on the user's system. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it ingests data from external RSS/Atom feeds.
- Ingestion points: Untrusted data enters the agent context via the
blogwatcher scancommand which fetches remote feed content (SKILL.md). - Boundary markers: Absent; there are no instructions to use delimiters or warnings to ignore instructions embedded in the feed data.
- Capability inventory: The skill uses subprocess calls to
blogwatcher(SKILL.md) and performs network operations. - Sanitization: Absent; the skill does not specify any validation or filtering for the content retrieved from external feeds.
Audit Metadata