hermes-cron

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides numerous examples and detailed references for using the hermes cron command-line interface to create, pause, resume, and monitor scheduled tasks.
  • [EXTERNAL_DOWNLOADS]: Instructions include checking for updates and reading release notes from the official repository at github.com/NousResearch/hermes-agent and documentation at hermes-agent.nousresearch.com.
  • [PERSISTENCE]: The core functionality described is the creation of scheduled, recurring jobs stored in the application's configuration (~/.hermes/cron/jobs.json), which allows tasks to persist and execute across sessions.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes recipes for 'watcher' jobs that ingest data from external sources or local workspace files (.hermes.md). This creates an attack surface where instructions embedded in those external sources could potentially influence the agent's behavior during a scheduled run.
  • Ingestion points: Jobs defined in templates/cron-jobs.md are designed to read from external <SOURCE> links or workspace context files.
  • Boundary markers: The templates do not explicitly include delimiters or 'ignore' instructions for the external content.
  • Capability inventory: Scheduled jobs can execute prompts with tool access, and specific 'script-only' jobs can execute shell scripts located in the $HERMES_HOME/scripts/ directory.
  • Sanitization: The skill does not define specific sanitization or filtering logic for the data ingested by these jobs.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 09:23 AM