skill-publish-analytics
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python analysis script (
scripts/analyze.py) and references a data export script (track.py) from a related built-in skill to prepare analytics data. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local logs (
publish-log.json) which contain fields like post titles and notes. These fields may contain content originating from external platforms, presenting a potential surface for indirect prompt injection when the agent interprets the analysis results. - Ingestion points: The skill reads
outputs/_analytics/publish-log.jsonandoutputs/_analytics/follower-log.jsonas specified inSKILL.mdand processed byscripts/analyze.py. - Boundary markers: Data is processed into a structured JSON summary by the analysis script before the agent reads it, but there are no specific markers or instructions to ignore embedded commands within the textual data fields.
- Capability inventory: The skill utilizes subprocess execution of local Python scripts as described in
SKILL.md. - Sanitization: There is no evidence of sanitization or escaping of untrusted textual content (e.g., titles, notes) before it is processed by the agent.
Audit Metadata