claude-recall@0.41.0
Persistent memory for Claude Code and Pi with native Skills integration, automatic capture, failure learning, and project scoping
AI Security Review
scanned 23d ago · by lpm-firewall-aiAutomatic install-time updates to existing package-owned agent hooks establish a concrete lifecycle risk. The inspected guards and backups support downgrade_to_warn rather than blocking for broad agent control hijacking.
Decision evidence
public snapshot- package.json enables automatic postinstall execution.
- scripts/postinstall.js invokes repair with automatic application across all scopes.
- The CLI forwards the automatic flag and selected scope to runRepair.
- Repair scans user-global settings and project settings beneath the home directory.
- Repairs target broken or incompatible existing Claude Recall hook commands, replacing them with the package command.
- Repair backs up settings before writing changes.
- The inspected repair preserves other hook entries and does not install hooks into configurations lacking them.
- Skill installation requires the explicit setup command; MCP consolidation requires the doctor fix option.
- No concrete credential theft or unrelated payload execution was established in the inspected chain.
Source & flagged code
20 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to claude-recall@0.41.0.
See all versions of claude-recall for other recorded verdicts.
- Review the evidence and your use of claude-recall@0.41.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.