SKILLEMALL.ai
GitHub ↗ RU Sign in

DC openweb

Typed JSON access to 90+ real websites — Google, Amazon, Reddit, YouTube, GitHub, Instagram, Bloomberg, Zillow, and more — across search, shopping, travel, finance, social, news, and dev tools. Use to read, search, post, comment, message, or otherwise interact with these sites (prices, products, articles, stock quotes, flights, posts, comments, DMs, etc.). Run `openweb sites` for the full list.

Not recommendedcritical or high security findings · low grade D
ClawHub Agent Skills author: imoonkey v0.1.5 MIT-0 23 files body ≈ 1 230 tokens Open the sourceclawhub.ai↗ analyzed 3 d ago

As a process C 53/100 · Has gaps — weak spots: result and completion, when it triggers, inputs and preconditions

IntegrationGitHubYouTubeInfrastructureWriting and documentsMarketingtype and topics are labelled automatically from the skill text
Runs in: OpenClaw Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) Goose OpenCode
JSON
Technical rating
D
57/100
safety, quality, tests
Safety 60%
36
Quality 40%
89
Run on models
none yet
Process rating
C
53/100
Has gaps
Result and completion w 14
0
Inputs and preconditions w 11
0
Progress reporting w 2
0
the three weakest of ten parameters · all ten

What is at stake

The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.

Exfiltration
If you install

The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".

For the author

If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.

How to improve

  1. Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
For the model run — optional
  • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
  • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

Guard findings · 5

  • high Exfiltration intent-browser-credential-store knowledge/bot-detection.md:61
    Accesses a browser credential / cookie store
    - **Warm-up requirement:** The browser warm-up phase must trigger all sensors. Wait for all vendor cookies to be set before executing operations.
  • high Exfiltration intent-browser-credential-store knowledge/bot-detection.md:258
    Accesses a browser credential / cookie store
    4. **The managed browser copies your real Chrome profile** -- it has history/cookies that look legitimate
  • high Exfiltration intent-browser-credential-store references/cli.md:56
    Accesses a browser credential / cookie store
    Auto-start copies auth-relevant files (Cookies, Local Storage, Session Storage, Web Data, Preferences) from the user's default Chrome profile to a temp directory, then launches Chrome with `--remote-d
Medium and low: 2
  • medium Exfiltration intent-browser-credential-store knowledge/auth-routing.md:99
    Accesses a browser credential / cookie store (detector / deny-list definition)
    The snapshot is now **copy-everything-minus-blocklist** (caches, `History`, `Sessions`, `Login Data`, locks, macOS `.com.google.Chrome.*` xattr sidecars). Marginally larger (~tens of MB), considerably
    detector
  • medium Exfiltration intent-browser-credential-store references/cli.md:202
    Accesses a browser credential / cookie store (quoted — discussed, not commanded)
    **Data boundary.** Cookies, localStorage, IndexedDB stay on the user's machine. Standard browser behavior sends cookies only to their originating sites. OpenWeb does not transmit any user data to open
    quoted

Files scanned: 23. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

✓ No remarks against the Agent Skills spec

Process rating: all ten parameters 53/100

  • 0Result and completion. Does not say what the result is
  • 0Inputs and preconditions. Does not say what the process needs to start
  • 0Progress reporting. Says nothing while it works
  • 20When it triggers. No condition that starts the skill
  • 30Running it twice. 2 mutating operations with no state check
  • 60Tools and files. Uses tools (node) that frontmatter does not declare
  • 100Steps. 6 steps
  • 100Failures and branches. 4 branches, has a failure section
  • 100Consistency. Name and required fields are in place
  • 100Execution cost. Instruction body is 1230 tokens
  • low The response is described with custom markup (7 tags): a typed call is more reliable

Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

Quality signals

  • +5Description has no quoted example phrases that should trigger the skill
  • +4Description does not say when NOT to use the skill (false activations)
  • +3Output format is not stated: the model decides each time
  • +2Single-language instructions
  • +3Description length 397: enough signal without eating the budget
  • +4Structure: 16 headings
  • +3Step-by-step instructions: 6 items
  • +4Has examples (4 code blocks)
  • +4Reference files are cited in the instructions (3 of 3)
  • +1License stated

Quality base 70; lint remarks subtract, signals add up to 100. Result: 89.

External checks

ClawHub: suspicious
OpenWeb is a coherent website-automation skill, but it needs review because it can reuse browser sessions and tokens to act on many real accounts.
LLM: suspicious (high) · VirusTotal: · 29 May 2026