daily-recap
Pass
Audited by Gen Agent Trust Hub on Mar 24, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes agent memory files and task summaries which may contain untrusted data from previous interactions.
- Ingestion points: Memory files (
memory/YYYY-MM-DD.md) and cron run summaries are read to identify accomplishments. - Boundary markers: None. The skill does not use delimiters to isolate the summarized content from the instructions.
- Capability inventory: The skill has the ability to read local files and invoke the
nano-banana-proimage generation tool. - Sanitization: Absent. The agent is instructed to select wins for the posterboard without any validation or escaping of the content.
- [DATA_EXFILTRATION]: The skill aggregates information from the agent's internal history and transmits a visual summary to an external destination defined by the
RECAP_CHAT_IDenvironment variable. - [COMMAND_EXECUTION]: Documentation provides a setup command for the user to manually create a persistent cron job using
clawdbot cron add.
Audit Metadata