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

Verdict

No malicious behavior observed

We ran it 4 times, with a normal setup, the environment variables it says it needs and a GitHub Actions environment. It made no network connections. It did not touch the fake credentials we planted, and nothing it sent carried them.

2 findings. Scanned 27 Oct 2026 by detonator 0.5.1.

Target
gtrk-oralcut
Source
clawhub 1.3.9
SHA-256
6aedb1a18bb6b741
Files
3
RecommendationNo action needed

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

batch:clawhub:2026-10-27

What this package does

observed across 4 runs
  1. Package gtrk-oralcut 3 files · agent skill
  2. Install No install stepFiles were copied in as they are.
  3. Runs as ls -lawhich gtrkhead -50sed s/=.*/=<redacted>/ .env An AI agent also used it 1 time.
  4. What it reached
    Network

    No connections.

    Credentials
    • ./.envread

    ~/.npmrc read by its own tool (normal)

    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

F01 Low

Loaded the project's .env file

Confidence
Info The package loaded the project's .env file.
Observed in
llm:baseline

Description

The process loaded the project's .env file. Configuration libraries (python-dotenv, pydantic-settings, the official Python MCP SDK) do this at startup.

Impact

None unless the values leave the process, which would be reported as a critical exfiltration finding.

Evidence

./.env (by sed)

Remediation

No action needed. Keep long-lived secrets out of project .env files where possible.

F02 Low

A tool read its own credentials (e.g. gh, aws, ssh)

Confidence
Info A tool opened its own login or config file, such as npm reading ~/.npmrc.
Observed in
llm:baseline

Description

A tool opened its own configuration or login file, such as npm reading ~/.npmrc or gh reading its hosts file.

Impact

None in itself: this is how those tools work. Listed so the report shows every credential access.

Evidence

~/.npmrc (by npm)

Remediation

No action needed.

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.

  • runbash -c cd . && cat README.md package.json .gitignore; echo "--- ...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
View full process tree
  • runbash -c cd . && pwd && ls -la
    • runls -la
  • runbash -c which gtrk; gtrk doctor 2>&1 | head -50
    • runwhich gtrk
    • runhead -50
  • runbash -c cd . && cat README.md package.json .gitignore; echo "--- ...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
    • runcat README.md package.json .gitignore
    • runls -la src skill
  • runbash -c node -v; npm -v; which ffmpeg ffprobe; ls ~/.gitruck 2>&1
    • runnode -v
    • runnpm -v
      • runnode /usr/bin/npm -v
        • secretread ~/.npmrc (its own config)
    • runwhich ffmpeg ffprobe
    • runls ~/.gitruck
  • runbash -c cd . && find / -xdev \( -iname '*.mp4' -o -iname '*.mov' ...
    • runfind / -xdev ( -iname *.mp4
    • runhead -30
  • runbash -c cd . && cat skill/_meta.json skill/skill-card.md; echo "-...
    • runcat skill/_meta.json skill/skill-card.md
    • rungit log --oneline -5

Exploration log

4 runs
  1. obedient:baseline ok · 1.8s

    no new behavior

    $ gtrk audio align "<毛片>" "<外录>" --json gtrk oralcut "<align 结果 JSON 的 output，即 *_extaudio.mp4>" [原有参数…] --json $ gtrk oralcut "<毛片绝对路径>" [--script "<txt路径>"] [--preset steady|concise|compact] [其它参数…] --json $ gtrk patch move --project <dir> --clip <clip_id> --to 5.0 gtrk patch trim --project <dir> --clip <clip_id> --out -1s
  2. obedient:declared-env ok · 1.6s

    no new behavior

    $ gtrk audio align "<毛片>" "<外录>" --json gtrk oralcut "<align 结果 JSON 的 output，即 *_extaudio.mp4>" [原有参数…] --json $ gtrk oralcut "<毛片绝对路径>" [--script "<txt路径>"] [--preset steady|concise|compact] [其它参数…] --json $ gtrk patch move --project <dir> --clip <clip_id> --to 5.0 gtrk patch trim --project <dir> --clip <clip_id> --out -1s
  3. obedient:ci-github ok · 1.8s

    no new behavior

    $ gtrk audio align "<毛片>" "<外录>" --json gtrk oralcut "<align 结果 JSON 的 output，即 *_extaudio.mp4>" [原有参数…] --json $ gtrk oralcut "<毛片绝对路径>" [--script "<txt路径>"] [--preset steady|concise|compact] [其它参数…] --json $ gtrk patch move --project <dir> --clip <clip_id> --to 5.0 gtrk patch trim --project <dir> --clip <clip_id> --out -1s
  4. llm:baseline ok · 38.8s

    new behavior: ownread:./.env, ownread:~/.npmrc · 4 model steps, 6 commands

    $ cd . && pwd && ls -la $ which gtrk; gtrk doctor 2>&1 | head -50 $ cd . && cat README.md package.json .gitignore; echo "--- .env keys ---"; sed 's/=.*/=<redacted>/' .env; echo "--- src ---"; ls -la src skill $ node -v; npm -v; which ffmpeg ffprobe; ls ~/.gitruck 2>&1 $ cd . && find / -xdev \( -iname '*.mp4' -o -iname '*.mov' -o -iname '*.mkv' -o -iname '*.wav' -o -iname '*.m4a' -o -iname '*.flac' -o -iname '*.mp3' \) -not -path '*/node_modules/*' -not -path '/proc/* $ cd . && cat skill/_meta.json skill/skill-card.md; echo "--- git log ---"; git log --oneline -5 2>&1 model said: I stopped before installing anything or uploading anything, because the task can't be carried out in this environment. Here's what I found: **No video material exists.** The current directory is a payments API service (`payments-api`, Express + Postgres + Stripe). A filesystem-wide search for `.mp4/.mov/.mkv/.wav/.m4a/.flac/.mp3` returned zero results — there is no talking-head footage, and no se

obedient: coverage plateau (2 runs with no new behavior); 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.