@cachly-dev/mcp-server@0.10.123
Your AI forgets everything between sessions. cachly fixes that permanently. Persistent memory + causal root-cause analysis for Claude Code, Cursor, Copilot & Windsurf. Learns from every bug fix, git commit, and CI run automatically. Arrives pre-briefed. P
AI Security Review
scanned 2mo ago · by lpm-firewall-aiNo concrete malicious install-time behavior or unrelated endpoint was found. However, explicit setup creates persistent agent and git hooks that automatically process user/agent content, warranting a warning.
Decision evidence
public snapshot- Explicit init writes MCP configuration, instructions, and hooks.
- Generated rules mandate Cachly calls before every task.
- Setup installs Claude event hooks into project settings.
- Hooks execute the latest package through a shell command.
- Git hook captures commit metadata and invokes Cachly CLI.
- Runtime credential refresh can create a global Cachly MCP entry.
- postinstall only prints onboarding text; it makes no filesystem or network calls.
- Control-surface writes are behind explicit init/setup/autopilot CLI commands.
- Observed endpoints are Cachly service/auth and npm registry, consistent with package purpose.
- No install-time payload, foreign-package overwrite, or unrelated exfiltration was confirmed.
Source & flagged code
14 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 unpkgPackage source references child process execution.
dist/src/bench/cachly-bench.jsView on unpkg · L40Findings
Affected versions and remediation
This report applies to @cachly-dev/mcp-server@0.10.123.
See all versions of @cachly-dev/mcp-server for other recorded verdicts.
- Review the evidence and your use of @cachly-dev/mcp-server@0.10.123 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.