meta-methodology-context-management
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: No malicious intent or dangerous patterns were identified. The skill is a methodology for project session persistence.
- [COMMAND_EXECUTION]: The skill uses
pwdandgit logto establish working context and review recent history. These are low-privilege discovery commands consistent with the skill's purpose. - [PROMPT_INJECTION]: Identifies an indirect prompt injection surface where the agent processes instructions and data from local documentation files.
- Ingestion points:
.claude/progress.md,.claude/decisions.md,.claude/insights.md,.claude/tests.json, and.claude/patterns.md(as defined in SKILL.md). - Boundary markers: None identified.
- Capability inventory: Filesystem read/write access and execution of
git logandpwd(found in SKILL.md). - Sanitization: None identified.
- [EXTERNAL_DOWNLOADS]: The skill's metadata references a schema from the vendor's official GitHub repository (
github.com/agents-inc/cli).
Audit Metadata