agent-daily-review
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE]: The skill scans and processes sensitive agent memory files (
MEMORY.mdand daily journal entries). While no network exfiltration was detected, the skill has full access to the agent's historical logs and workspace artifacts. - [PERSISTENCE_MECHANISMS]: The skill provides instructions and examples for setting up persistence using
openclaw cronor standard system cron jobs to trigger the review script daily. This is a documented core feature for growth tracking. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from daily journal entries and artifacts and incorporates them into reports without sanitization.
- Ingestion points:
memory/YYYY-MM-DD.md,MEMORY.md, andworkspace/*.md(found inscripts/daily_review.py). - Boundary markers: Absent; there are no specific delimiters to prevent the agent from following instructions embedded in journals.
- Capability inventory: The script performs local file reading and writing within the workspace.
- Sanitization: Absent; content is extracted via regex and keyword matching and appended directly to Markdown reports.
Audit Metadata