ccs-align
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses local commands such as bun test for regression testing and ripgrep (rg) for verifying plan documentation. These are standard development operations scoped to the local repository.
- [DYNAMIC_EXECUTION]: A Node.js one-liner is employed to dynamically resolve the local worker port by reading from ~/.claude-mem/settings.json. This is a routine configuration task for the application environment.
- [DATA_EXFILTRATION]: While the skill reads and writes data, all operations are confined to the local machine between the claude-mem worker API (127.0.0.1) and a local cache directory (~/.claude-mem/ccs-align/). No data exfiltration to external domains was detected.
- [PROMPT_INJECTION]: The skill includes instructions to address the user as Alex and follow specific persona constraints. These are functional requirements for agent behavior and do not constitute malicious injection or safety bypasses.
- [SAFE]: The skill belongs to the official claude-mem repository ecosystem authored by thedotmack. It implements safety best practices including atomic file writes and environment-gated patching of rule conflicts.
Audit Metadata