life-replan
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied input to describe why a plan was invalidated and records this in a project log. While this represents a data ingestion surface, the risk is mitigated by the 'secretary' role defined in the referenced
ROLE.md, which requires user confirmation before applying material changes and restricts the agent's capabilities to note-taking and vault management. - Ingestion points:
SKILL.md(Step 1: 'State what invalidated the plan (user words)'). - Boundary markers: Absent in the prompt templates.
- Capability inventory: Appends text to project logs and modifies project metadata (
next_action) inSKILL.md. - Sanitization: Requires explicit user confirmation for proposed changes as specified in
ROLE.md('wait for confirm on material state') andSKILL.md('apply after confirm'). - [SAFE]: The skill uses a sibling configuration file (
ROLE.md) to establish a 'secretary' persona, which acts as a security guardrail by explicitly forbidding autonomous project status changes and external file modifications.
Audit Metadata