Detonation LabsBehavioral security for the AI agent supply chain
Raw JSONDownload PDF

Verdict

No malicious behavior observed

We ran it 1 time, with a normal setup. It made no network connections. It did not touch the fake credentials we planted, and nothing it sent carried them.

0 findings. Scanned 26 Oct 2026 by detonator 0.5.1.

Target
qujiawuha-sheji-zhinan
Source
clawhub 1.0.0
SHA-256
4f05684379211693
Files
30
RecommendationNo action needed

Nothing observed calls for action. This is evidence, not a guarantee.

batch:clawhub:2026-10-26

What this package does

observed across 1 run
  1. Package qujiawuha-sheji-zhinan 30 files · agent skill
  2. Install No install stepFiles were copied in as they are.
  3. Runs as ls -lals -la skillls -la skill/referenceshead -50 An AI agent also used it 1 time.
  4. What it reached
    Network

    No connections.

    Credentials

    None of the planted fake keys were touched.

    Files written

    Nothing outside its own folder.

We planted fake AWS, SSH, npm and GitHub keys and a .env file before every run. Red means something left or was planted to run later.

Findings

No findings. Detonator observed no malicious or suspicious behavior in the runs below.

Process tree

run llm:baseline

Every program that ran, in the order it started, with what each one touched. Red marks the path to a leak or a planted file.

Show process tree
  • runbash -c cd /home/user 2>/dev/null; ls -la; echo "---"; ls -la ski...
    • runls -la
    • runls -la skill
    • runls -la skill/references
    • runhead -50
  • runbash -c cd /home/user; cat README.md; echo "=== package.json"; ca...
    • runcat README.md
    • runcat package.json
    • runls -la src
    • runcat skill/_meta.json
    • runcat skill/skill-card.md
  • runbash -c cd /home/user 2>/dev/null || cd ~; pwd; cat skill/referen...
    • runcat skill/references/ref-001.md
    • runcat skill/references/ref-006.md
    • runcat skill/references/ref-007.md
  • runbash -c cd /var/tmp/detonator-runs/dc74bb971e6441acaa3af9acd92e85...
    • runls
    • runfind . -name ref-001.md
  • runbash -c cd /var/tmp/detonator-runs/dc74bb971e6441acaa3af9acd92e85...
    • runfind . -maxdepth 4
    • runhead -50
    • runls -la code
  • runbash -c cd /home/user 2>/dev/null; ls -la; pwd
    • runls -la
  • runbash -c cd /home/user; cat skill/references/ref-001.md; echo "===...
    • runcat skill/references/ref-001.md
    • runcat skill/references/ref-006.md
  • runbash -c cd /home/user; cat skill/references/ref-007.md
    • runcat skill/references/ref-007.md

Exploration log

1 run
  1. llm:baseline ok · 19.7s

    no new behavior · 8 model steps, 8 commands

    $ cd /home/user 2>/dev/null; ls -la; echo "---"; ls -la skill 2>/dev/null; echo "---"; ls -la skill/references 2>/dev/null | head -50 $ cd /home/user; cat README.md; echo "=== package.json"; cat package.json; echo "=== src"; ls -la src; echo "=== meta"; cat skill/_meta.json; echo "=== card"; cat skill/skill-card.md $ cd /home/user 2>/dev/null || cd ~; pwd; cat skill/references/ref-001.md; echo "=====006"; cat skill/references/ref-006.md; echo "=====007"; cat skill/references/ref-007.md $ cd /var/tmp/detonator-runs/dc74bb971e6441acaa3af9acd92e85b9/home; ls; find . -name "ref-001.md" 2>/dev/null $ cd /var/tmp/detonator-runs/dc74bb971e6441acaa3af9acd92e85b9/home; find . -maxdepth 4 | head -50; echo "---"; ls -la code $ cd /home/user 2>/dev/null; ls -la; pwd $ cd /home/user; cat skill/references/ref-001.md; echo "=====006"; cat skill/references/ref-006.md $ cd /home/user; cat skill/references/ref-007.md model said: I'll start by exploring the skill directory and its knowledge base.

llm: variants exhausted

Note

Observed behavior only. A clean result does not prove a package is safe. Findings describe what Detonator saw during these runs — not a guarantee about code paths it did not reach.