Waxell

plugin

review-agent-governance High risk

wshobsonwshobson/agentsupdated

Require a human approval signal before an AI agent can post PR reviews, comments, merges, or writes to CI config. Cedar-gated, receipt-signed, designed for the Hermes-style failure mode where a review bot posts without oversight.
wshobson's own description, from the plugin's files.
Stars
42,753
Version
v3
Copies
0

Overview

blockHigh risk

Waxell rates it high risk: 18 findings, highest severity high.

Highest severity
high
Findings
18
What ran
Static analysis + deep scan

A clean result means nothing alarming in what is published, not a guarantee. How the checks work

We don't republish the plugin's instructions. Read them on GitHub at the reviewed commit.

Security

Security scan

Checked 2 days ago

Result

High risk Waxell rates it high risk: 18 findings, highest severity high.

Checked with static analysis + deep scan: servers it launches, packages it installs at runtime, scripts piped to a shell and bundled binaries. The deep scan also read the instructions for prompt injection and data exfiltration. Nothing was run. How the checks work

Findings (18)

high · 10

  • high

    Tool Misuse: Tool Parameter Abuse

    README.md:126 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    README.md:200 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    commands/approve-review.md:80 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    skills/review-agent-setup/SKILL.md:81 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    skills/review-agent-setup/SKILL.md:153 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    commands/approve-review.md:10 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    commands/approve-review.md:25 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    commands/approve-review.md:107 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    commands/approve-review.md:96 · skillspector:TM1

  • high

    Tool Misuse: Tool Parameter Abuse

    test/run-tests.sh:122 · skillspector:TM1

medium · 8

  • medium

    Excessive Agency: Autonomous Decision Making

    README.md:180 · skillspector:EA2

  • medium

    Privilege Escalation: Sudo/Root Execution

    README.md:263 · skillspector:PE2

  • medium

    Privilege Escalation: Sudo/Root Execution

    skills/review-agent-setup/SKILL.md:57 · skillspector:PE2

  • medium

    Excessive Agency: Autonomous Decision Making

    commands/list-pending.md:47 · skillspector:EA2

  • medium

    Excessive Agency: Autonomous Decision Making

    skills/review-agent-setup/SKILL.md:97 · skillspector:EA2

  • medium

    npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and pub

    hooks/evaluate.sh:43 · skillspector:RP1

  • medium

    Bundled lifecycle hooks can run automatically when their configured events occur, so their reach and handler capability

    hooks/hooks.json:1 · skillspector:BH1

  • medium

    Excessive Agency: Autonomous Decision Making

    test/run-tests.sh:111 · skillspector:EA2

Versions

Version history

Each distinct content we have seen, newest first, with the verdict it got at the time.

  1. v3 (0.2.0)2026-11-196 changed High riskgot riskier
  2. v2 (0.1.6)2026-11-166 changed No issues
  3. v1 (0.1.5)2026-11-03 No issues

Files

Files

16 files, 61 KB in total. Names only; open them on GitHub.

  • .claude-plugin/plugin.json
  • .codex-plugin/plugin.json
  • README.md
  • agents/review-policy-author.md
  • commands/approve-review.md
  • commands/list-pending.md
  • hooks/evaluate.sh
  • hooks/hooks.json
  • hooks/sign.sh
  • policies/review-agent-governance.cedar
  • policies/review-agent-governance.cedarschema
  • skills/review-agent-setup/SKILL.md
  • test/README.md
  • test/fixtures/pretool-allow-read.json
  • test/fixtures/pretool-deny-bash-merge.json
  • test/run-tests.sh

See and control which skills your people use

Inside your organization, Waxell scans each skill and plugin before anyone installs it, approves what passes and publishes it to a private marketplace, and puts every MCP tool call behind one governed URL.

30 minutes. No commitment.

SOC 2 Type II attested · ISO/IEC 27001:2022 certified · HIPAA with a BAA · aligned to the EU AI Act and NIST AI RMF

More design skills and plugins

See all