registry / deweyou-cli / 1.1.0

deweyou-cli@1.1.0

Dewey's personal agent workflow bootstrapper

AI Security Review

scanned 16h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. This is an explicit agent-workflow bootstrapper that can install remote Dewey assets and write selected Codex/Claude configuration. No unconsented install-time control-surface mutation or exfiltration is established.

Static reason

No blocking static signals were detected.

Trigger

User runs `deweyou-cli agent update` and then `agent init`, especially with global scope.

Impact

Selected remote content can influence Codex/Claude behavior through user-approved configuration files.

Mechanism

Remote asset cache plus user-invoked agent instruction/skill installation.

Rationale

The package has a real, user-invoked capability to alter AI-agent configuration using remotely sourced assets. Its lifecycle script only builds the package, while explicit initialization includes selection/confirmation and write-safety checks.

Evidence

package.jsondist/deweyou.mjsdist/cache-CSIIRn48.mjsdist/init-B1M_gTEI.mjsdist/dev-DEZzOk5W.mjsdist/prompts-Bz0ZqiO8.mjs$HOME/.codex/AGENTS.md$HOME/.claude/CLAUDE.md$HOME/.codex/hooks.json.agents/manifest.jsonAGENTS.md

Network endpoints1

github.com/deweyou/agents.git

deweyou-cli@1.1.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code