registry / @askalf/truecopy / 0.9.0

@askalf/truecopy@0.9.0

own your agent skills — vet, sign, and pin every skill & MCP server before it runs. The supply-chain gate for AI agents. Part of Own Your Stack.

AI Security Review

scanned 1h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. No install-time attack behavior is present. An explicit user command can persist a truecopy-owned Claude Skill gate; its default command later uses an npx GitHub package reference.

Static reason

One or more suspicious static signals were detected.

Trigger

User runs `truecopy hook install` or `truecopy hook install --user`.

Impact

Persists a package-owned agent hook that participates in future Skill invocations; no unconsented install-time mutation or exfiltration is established.

Mechanism

Explicit Claude PreToolUse hook installation and caller-directed MCP process wrapper.

Rationale

Source inspection supports a warning for explicit persistent agent-hook setup under the install-control-surface policy, not a malware block. The scanner's environment-variable and reviewer-text hints do not establish malicious behavior.

Evidence

package.jsonsrc/cli.mjssrc/mcp-cli.mjssrc/mcp.mjssrc/scan.mjssrc/keychain.mjs.claude/settings.json~/.claude/settings.jsontruecopy.lockcanon.lock~/.canon/trust.json

Decision evidence

public snapshot

AI called this at 94.0% confidence with false-positive risk.

@askalf/truecopy@0.9.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code