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 a production environment. It contacted registry.npmjs.org. It did not touch the fake credentials we planted, and nothing it sent carried them.

4 findings. Scanned 27 Oct 2026 by detonator 0.5.1.

Target
mermaid-chart
Source
clawhub 1.0.0
SHA-256
58894a9bed67a845
Files
4
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 mermaid-chart 4 files · agent skill
  2. Install No install scripts of its ownOnly the package manager ran.
  3. Runs as ls -lals -R ./skillhead -50sed s/=.*/=<redacted>/ .env An AI agent also used it 1 time.
  4. What it reached
    Network
    • registry.npmjs.org 5 connections, 1.6 KB sent
    Credentials
    • ./.envread

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

    Files written
    • ./B:written
    • ./Cancelledwritten
    • ./Check{Validation Passed?}written
    • ./Completedwritten
    • ./D:written
    • ./End[End]written
    • ./F:written
    • ./Nowritten
    • ./Paidwritten
    • ./Pendingwritten
    • ./PendingPaymentwritten
    • ./Processingwritten

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

Used a planted key with the service it belongs to (registry.npmjs.org)

Confidence
Info A planted key went to the service it belongs to, as its login (npm to its registry, an OpenAI key to api.openai.com).
Observed in
llm:baseline

Description

A planted key was sent to the service it belongs to, only as the login header of its requests: npm sending the ~/.npmrc token to registry.npmjs.org while installing, or an app calling api.openai.com with the OpenAI key from .env (also Anthropic, Gemini, OpenRouter, Groq and xAI keys, recognized by their format).

Impact

Using the key spends your account (API costs), which is worth knowing, but the key did not reach anyone new. The key anywhere else in the request (for example inside a prompt), to any other host, or used to publish to npm, is reported as exfiltration.

Evidence

~/.npmrc (npm) sent only as the login header of requests to registry.npmjs.org

Remediation

No action needed.

F02 Low

Network destinations contacted

Confidence
Info The package connected to these hosts.
Observed in
llm:baseline

Description

Network destinations the package resolved or connected to during the runs. All traffic was captured by the sandbox; nothing reached the internet.

Impact

Context only. Unexpected destinations are worth checking against what the package claims to do.

Evidence

registry.npmjs.org

Remediation

Confirm each host is expected (the package's API, its registry). Allowlist expected egress in production.

F03 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.

F04 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 cat README.md; echo "--- package.json"; cat package.json;...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
View full process tree
  • runbash -c ls -la; echo "---"; ls -R ./skill 2>/dev/null | head -50
    • runls -la
    • runls -R ./skill
    • runhead -50
  • runbash -c cat README.md; echo "--- package.json"; cat package.json;...
    • runsed s/=.*/=<redacted>/ .env
      • secretread ./.env
    • runcat README.md
    • runcat package.json
    • runls -la src
    • rungit log --oneline
    • runhead -20
  • runbash -c cat skill/skill-card.md; echo "=== _meta"; cat skill/_met...
    • runcat skill/skill-card.md
    • runcat skill/_meta.json
    • runwc -l skill/references/syntax-reference.md
  • runbash -c cat src/index.js; echo "=== ref"; cat skill/references/sy...
    • runcat src/index.js
    • runcat skill/references/syntax-reference.md
  • runbash -c cat -A src/index.js | head -20; echo "=== size"; wc -c sr...
    • runcat -A src/index.js
    • runhead -20
    • runwc -c src/index.js
    • runhead -60 skill/references/syntax-reference.md
  • runbash -c which node npm npx mmdc 2>/dev/null; node -v 2>/dev/null;...
    • runwhich node npm npx mmdc
    • runnode -v
    • runls node_modules
    • runhead
    • runtimeout 20 npm view mermaid version
      • runnpm view mermaid version
        • runnode /usr/bin/npm view mermaid version
          • secretread ~/.npmrc (its own config)
          • netregistry.npmjs.org:443 sent 433 B
          • netregistry.npmjs.org:443 sent 385 B
    • runtail -3
  • runbash -c cd /tmp && timeout 60 npm view mermaid version 2>&1 | tai...
    • runtail -5
    • runtimeout 60 npm view mermaid version
      • runnpm view mermaid version
        • runnode /usr/bin/npm view mermaid version
          • secretread ~/.npmrc (its own config)
          • netregistry.npmjs.org:443 sent 385 B
  • runbash -c cd /tmp && timeout 90 npm install mermaid@11 --no-audit -...
    • runtimeout 90 npm install mermaid@11 --no-audit
      • runnpm install mermaid@11 --no-audit --no-fund
        • runnode /usr/bin/npm install mermaid@11 --no-audit --no-fund
          • secretread ~/.npmrc (its own config)
          • netregistry.npmjs.org:443 sent 388 B
    • runtail -5
    • runhead
    • runls /tmp/node_modules/mermaid/dist
  • runbash -c cd /tmp && timeout 30 npm config get registry; timeout 30...
    • runtimeout 30 npm config get registry
      • runnpm config get registry
        • runnode /usr/bin/npm config get registry
          • secretread ~/.npmrc (its own config)
    • runtimeout 30 curl -sS -o /dev/null
      • runcurl -sS -o /dev/null -w %{http_code}\n
        • netregistry.npmjs.org:443 sent 88 B
        • netDNS lookups for npmjs.org 10 queries

Exploration log

4 runs
  1. obedient:baseline ok · 3.6s

    new behavior: write:./B:, write:./Cancelled, write:./Check{Validation Passed?}, write:./Completed

    $ flowchart TD Start["Start"] --> Input[/"Input Data"/] Input --> Check{"Validation Passed?"} Check -->|Yes| Process["Process Data"] Check -->|No| Error["Error"] Process --> Output[/ $ ["Rectangle"] (/Parallelogram/) ["Rounded Rectangle"] (("Circle")) {"Diamond"} [["Subroutine"]] [("Database")] >"Tag"] ((("Double Circle")) $ sequenceDiagram actor U as User participant F as Frontend participant B as Backend participant D as Database U->>F: Click Login F->>B: POST /api/login B->>D: Query User $ gantt title Project Schedule dateFormat YYYY-MM-DD section Design Requirements Analysis :done, req, 2026-01-06, 5d UI Design :active, ui, after req, 7d sectio $ classDiagram class Animal { +String name +int age +makeSound() void } class Dog { +String breed +fetch() void } class Cat { +String $ stateDiagram-v2 [*] --> PendingPayment PendingPayment --> Paid : Payment Successful PendingPayment --> Cancelled : Timeout/Manual Cancel Paid --> Shipped : Warehouse Dispatch Shipp $ erDiagram USER { int id PK string name string email UK } ORDER { int id PK int userId FK float amount datetime createdAt } U $ pie title User Source Distribution "Search Engine" : 45.2 "Direct Visit" : 28.7 "Social Media" : 15.3 "Email Marketing" : 10.8 $ mindmap root((Project Architecture)) Frontend React State Management UI Component Library Backend API Gateway Microservices User Service Order S $ timeline title Project Milestones 2024-Q1 : Project Launch : Tech stack finalized 2024-Q2 : MVP Development : Core features live 2024-Q3 : Beta Testing : Collect user feedback 2024
  2. obedient:ci-github ok · 3.3s

    no new behavior

    $ flowchart TD Start["Start"] --> Input[/"Input Data"/] Input --> Check{"Validation Passed?"} Check -->|Yes| Process["Process Data"] Check -->|No| Error["Error"] Process --> Output[/ $ ["Rectangle"] (/Parallelogram/) ["Rounded Rectangle"] (("Circle")) {"Diamond"} [["Subroutine"]] [("Database")] >"Tag"] ((("Double Circle")) $ sequenceDiagram actor U as User participant F as Frontend participant B as Backend participant D as Database U->>F: Click Login F->>B: POST /api/login B->>D: Query User $ gantt title Project Schedule dateFormat YYYY-MM-DD section Design Requirements Analysis :done, req, 2026-01-06, 5d UI Design :active, ui, after req, 7d sectio $ classDiagram class Animal { +String name +int age +makeSound() void } class Dog { +String breed +fetch() void } class Cat { +String $ stateDiagram-v2 [*] --> PendingPayment PendingPayment --> Paid : Payment Successful PendingPayment --> Cancelled : Timeout/Manual Cancel Paid --> Shipped : Warehouse Dispatch Shipp $ erDiagram USER { int id PK string name string email UK } ORDER { int id PK int userId FK float amount datetime createdAt } U $ pie title User Source Distribution "Search Engine" : 45.2 "Direct Visit" : 28.7 "Social Media" : 15.3 "Email Marketing" : 10.8 $ mindmap root((Project Architecture)) Frontend React State Management UI Component Library Backend API Gateway Microservices User Service Order S $ timeline title Project Milestones 2024-Q1 : Project Launch : Tech stack finalized 2024-Q2 : MVP Development : Core features live 2024-Q3 : Beta Testing : Collect user feedback 2024
  3. obedient:production ok · 4.2s

    no new behavior

    $ flowchart TD Start["Start"] --> Input[/"Input Data"/] Input --> Check{"Validation Passed?"} Check -->|Yes| Process["Process Data"] Check -->|No| Error["Error"] Process --> Output[/ $ ["Rectangle"] (/Parallelogram/) ["Rounded Rectangle"] (("Circle")) {"Diamond"} [["Subroutine"]] [("Database")] >"Tag"] ((("Double Circle")) $ sequenceDiagram actor U as User participant F as Frontend participant B as Backend participant D as Database U->>F: Click Login F->>B: POST /api/login B->>D: Query User $ gantt title Project Schedule dateFormat YYYY-MM-DD section Design Requirements Analysis :done, req, 2026-01-06, 5d UI Design :active, ui, after req, 7d sectio $ classDiagram class Animal { +String name +int age +makeSound() void } class Dog { +String breed +fetch() void } class Cat { +String $ stateDiagram-v2 [*] --> PendingPayment PendingPayment --> Paid : Payment Successful PendingPayment --> Cancelled : Timeout/Manual Cancel Paid --> Shipped : Warehouse Dispatch Shipp $ erDiagram USER { int id PK string name string email UK } ORDER { int id PK int userId FK float amount datetime createdAt } U $ pie title User Source Distribution "Search Engine" : 45.2 "Direct Visit" : 28.7 "Social Media" : 15.3 "Email Marketing" : 10.8 $ mindmap root((Project Architecture)) Frontend React State Management UI Component Library Backend API Gateway Microservices User Service Order S $ timeline title Project Milestones 2024-Q1 : Project Launch : Tech stack finalized 2024-Q2 : MVP Development : Core features live 2024-Q3 : Beta Testing : Collect user feedback 2024
  4. llm:baseline ok · 41.1s

    new behavior: dns:registry.npmjs.org, exec:curl, ownread:./.env, ownread:~/.npmrc · 8 model steps, 9 commands

    $ ls -la; echo "---"; ls -R ./skill 2>/dev/null | head -50 $ cat README.md; echo "--- package.json"; cat package.json; echo "--- src"; ls -la src; echo "--- .env (keys only)"; sed 's/=.*/=<redacted>/' .env; echo "--- git log"; git log --oneline | head -20 $ cat skill/skill-card.md; echo "=== _meta"; cat skill/_meta.json; echo "=== ref"; wc -l skill/references/syntax-reference.md $ cat src/index.js; echo "=== ref"; cat skill/references/syntax-reference.md $ cat -A src/index.js | head -20; echo "=== size"; wc -c src/index.js; echo "=== head of ref"; head -60 skill/references/syntax-reference.md $ which node npm npx mmdc 2>/dev/null; node -v 2>/dev/null; ls node_modules 2>/dev/null | head; timeout 20 npm view mermaid version 2>&1 | tail -3 $ cd /tmp && timeout 60 npm view mermaid version 2>&1 | tail -5; echo "exit=$?" $ cd /tmp && timeout 90 npm install mermaid@11 --no-audit --no-fund 2>&1 | tail -5; ls /tmp/node_modules/mermaid/dist 2>/dev/null | head $ cd /tmp && timeout 30 npm config get registry; timeout 30 curl -sS -o /dev/null -w "%{http_code}\n" https://registry.npmjs.org/mermaid 2>&1 model said: I'll start by checking the project directory and the skill's reference files.

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.