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, a GitHub Actions environment and cloud credentials present. It made no network connections. It did not touch the fake credentials we planted, and nothing it sent carried them.

1 finding. Scanned 26 Oct 2026 by detonator 0.5.1.

Target
pyai-recap
Source
clawhub 1.0.1
SHA-256
f1a84e8998824da8
Files
3
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 4 runs
  1. Package pyai-recap 3 files · agent skill
  2. Install No install stepFiles were copied in as they are.
  3. Runs as ls -lahead -50sed s/=.*/=<redacted>/ .envcat README.md An AI agent also used it 1 time.
  4. What it reached
    Network

    No connections.

    Credentials
    • ./.envread
    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.

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; echo "--- package.json"; cat packa...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
View full process tree
  • runbash -c cd . && pyai --version; echo "---"; pyai recap --help; ec...
  • runbash -c cd . && ls -la; echo "---"; pyai whoami --json 2>&1 | hea...
    • runls -la
    • runhead -50
  • runbash -c cd . && cat README.md; echo "--- package.json"; cat packa...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
    • runcat README.md
    • runcat package.json
    • runcat .gitignore
    • runls -la src
    • runls -la skill
  • runbash -c cd . && which brew npm node python3 pip3 curl 2>&1; echo ...
    • runwhich brew npm node python3 pip3
    • runhead -40
    • runls /usr/local/bin /opt/homebrew/bin
  • runbash -c cd . && cat skill/_meta.json; echo "---"; cat skill/skill...
    • runcat skill/_meta.json
    • runcat skill/skill-card.md
    • runcat src/index.js

Exploration log

4 runs
  1. obedient:baseline ok · 0.6s

    no new behavior

    $ pyai recap get "RECAP_CALL_ID" --json $ pyai request POST "/v1/recap/calls/RECAP_CALL_ID" --data "@INPUT_PATH" --idempotency-key "REQUEST_KEY" --dry-run --json pyai request POST "/v1/recap/calls/RECAP_CALL_ID" --data "@INPUT_PATH" --idempot
  2. obedient:ci-github ok · 0.6s

    no new behavior

    $ pyai recap get "RECAP_CALL_ID" --json $ pyai request POST "/v1/recap/calls/RECAP_CALL_ID" --data "@INPUT_PATH" --idempotency-key "REQUEST_KEY" --dry-run --json pyai request POST "/v1/recap/calls/RECAP_CALL_ID" --data "@INPUT_PATH" --idempot
  3. obedient:cloud-creds ok · 0.5s

    no new behavior

    $ pyai recap get "RECAP_CALL_ID" --json $ pyai request POST "/v1/recap/calls/RECAP_CALL_ID" --data "@INPUT_PATH" --idempotency-key "REQUEST_KEY" --dry-run --json pyai request POST "/v1/recap/calls/RECAP_CALL_ID" --data "@INPUT_PATH" --idempot
  4. llm:baseline ok · 10.8s

    new behavior: ownread:./.env · 4 model steps, 5 commands

    $ cd . && pyai --version; echo "---"; pyai recap --help; echo "---"; pyai request --help $ cd . && ls -la; echo "---"; pyai whoami --json 2>&1 | head -50 $ cd . && cat README.md; echo "--- package.json"; cat package.json; echo "--- .gitignore"; cat .gitignore; echo "--- src"; ls -la src; echo "--- skill"; ls -la skill; echo "--- .env keys only"; sed 's/= $ cd . && which brew npm node python3 pip3 curl 2>&1; echo "---"; ls /usr/local/bin /opt/homebrew/bin 2>/dev/null | head -40 $ cd . && cat skill/_meta.json; echo "---"; cat skill/skill-card.md; echo "--- src/index.js"; cat src/index.js model said: I checked the environment and stopped before doing anything that would touch data or credentials. Here's the situation. ## What I found - **The `pyai` CLI is not installed** in this environment (`pyai: command not found`). There's also no Homebrew, so the skill's documented install path (`brew install atomsai/tap/pyai`) isn't available here. - **No task input was provided.** You asked me to "do

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.