CC veritier

Agent Trust Gate for tool calls. Agents should not act without a gate and evidence. attest_action looks the action up in a system of record - npm, PyPI, CourtListener, SEC EDGAR, the U.S. Code, or a policy corpus you supply - and blocks the call if the record isn't there. Each decision is a reconstructable, revocable Ed25519-signed workpaper (Claim Credential), not a certificate of truth. Also extracts and verifies claims and scans documents. Connects over MCP Streamable HTTP, so there is nothing to install.

ClawHub Agent Skills author: Salomon Henao v3.4.2 MIT-0 43 files body ≈ 8 033 tokens Open the sourceclawhub.ai↗ analyzed 2 d ago

Agent Trust Gate for tool calls.

As a process C 63/100 · Has gaps — weak spots: inputs and preconditions, consistency, execution cost

Procedure AI and agentstype 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

68/100

safety, quality, tests

Safety 60%: 63
Quality 40%: 75
Run on models: none yet

Process rating

63/100

Has gaps

Inputs and preconditions w 11: 0
Running it twice w 4: 30
Consistency w 8: 40

the three weakest of ten parameters · all ten

What is at stake

Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.

Exfiltration medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

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

The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.

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 · 21

✓ No critical or high findings

Medium and low: 21

medium Exfiltration net-redirectable-api-key javascript/quickstart/test_integration.mjs:15

Helper sends the API key to a host configured by an environment variable — the key can be redirected to another server
```
API key + configurable base URL from environment
```
medium Exfiltration net-credential-use javascript/use-cases/informational_credential.mjs:55

Credential used in a network call (verify the destination is the intended service)
```
const pub = await fetch(`${API_URL}/v1/credentials/${cred.id}`);
```
medium Exfiltration exfil-send-secrets-to-url SKILL.md:56

Instruction to send secrets/history to an external endpoint ("only send to …" — scoping, not exfiltration; documentation of a security skill; the skill's own vendor host)
```
🔒 **SECURITY:** Only send your API key to `https://api.veritier.ai` - never to any other domain.
```
scopedsecurity skillvendor-host
medium Exfiltration exfil-send-secrets-to-url SKILL.md:590

Instruction to send secrets/history to an external endpoint ("only send to …" — scoping, not exfiltration; documentation of a security skill; the skill's own vendor host)
```
- Only send your API key to `https://api.veritier.ai`
```
scopedsecurity skillvendor-host
low Secrets in code secret-high-entropy-token javascript/package-lock.json:22

High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
```
"integrity": "sha5…GLi+2W/6ao+6Y7gu/RCwR…Kng==",
```
detector
low Secrets in code secret-high-entropy-token javascript/package-lock.json:137

High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
```
"integrity": "sha5…GLw+xYSd…cqA==",
```
detector
low Secrets in code secret-high-entropy-token javascript/package-lock.json:204

High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
```
"integrity": "sha5…FrF+LTRo…W3g==",
```
detector
low Secrets in code secret-high-entropy-token javascript/package-lock.json:213

High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
```
"integrity": "sha5…A6j+hAmM…GbS+kf5c…csw==",
```
detector
low Secrets in code secret-high-entropy-token javascript/package-lock.json:222

High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
```
"integrity": "sha5…5bm+c2gQ…aG5+esrLODihIorn+Pe6F…dXA==",
```
detector
low Exfiltration read-dotenv javascript/quickstart/extract_text.mjs:9

Reads a .env file (code comment)
```
*   2. cp .env.example .env  (then add your API key)
```
comment
low Exfiltration net-credential-use javascript/quickstart/reconstruct_credential.mjs:34

Credential used in a network call (verify the destination is the intended service) (documentation of a security skill)
```
const pub = await fetch(`${API_URL}/v1/credentials/${credId}`);
```
security skill
low Exfiltration net-credential-use javascript/quickstart/reconstruct_credential.mjs:61

Credential used in a network call (verify the destination is the intended service) (documentation of a security skill)
```
const revoked = await fetch(`${API_URL}/v1/credentials/${credId}/revoke`, {
```
security skill
low Exfiltration read-dotenv javascript/quickstart/verify_text.mjs:9

Reads a .env file (code comment)
```
*   2. cp .env.example .env  (then add your API key)
```
comment
low Exfiltration read-dotenv javascript/README.md:16

Reads a .env file (documentation of a security skill)
```
cp ../.env.example .env
```
security skill
low Exfiltration read-dotenv python/quickstart/attest_action.py:15

Reads a .env file (test fixture / example file; quoted — discussed, not commanded)
```
2. cp .env.example .env  (then add your API key)
```
fixturequoted
low Exfiltration read-dotenv python/quickstart/extract_text.py:10

Reads a .env file (quoted — discussed, not commanded)
```
2. cp .env.example .env  (then add your API key)
```
quoted
low Exfiltration read-dotenv python/quickstart/test_integration.py:11

Reads a .env file (test fixture / example file; quoted — discussed, not commanded)
```
3. cp .env.example .env  and set VERITIER_TEST_KEY=vt_test_...
```
fixturequoted
low Exfiltration read-dotenv python/quickstart/validate_document.py:9

Reads a .env file (quoted — discussed, not commanded)
```
2. cp .env.example .env  (then add your API key)
```
quoted
low Exfiltration read-dotenv python/quickstart/verify_text.py:9

Reads a .env file (quoted — discussed, not commanded)
```
2. cp .env.example .env  (then add your API key)
```
quoted
low Exfiltration read-dotenv python/README.md:16

Reads a .env file (documentation of a security skill)
```
cp ../.env.example .env
```
security skill
low Exfiltration read-dotenv README.md:215

Reads a .env file (documentation of a security skill)
```
cp .env.example .env
```
security skill

Files scanned: 42. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

warning body-long SKILL.md body ≈ 8033 tokens (recommended < 5000); move details to references/
note frontmatter-key unknown frontmatter key "homepage"

Process rating: all ten parameters 63/100

0Inputs and preconditions. Does not say what the process needs to start
30Running it twice. 20 mutating operations with no state check
40Consistency. Frontmatter name (veritier) differs from the folder (veritier-fact-checking)
40Execution cost. Instruction body is 8033 tokens: crowds the task out of the window
60Tools and files. Uses tools (web, python) that frontmatter does not declare
60Result and completion. Output format stated, no completion criterion
85Steps. 44 steps, 1 vague phrases
100When it triggers. States when to use and when not to
100Failures and branches. 1 branches, has a failure section
100Progress reporting. Reports progress

medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
low 11 top-level sections: this looks like several domains in one skill

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)
-229 emoji in the instructions: noise for the model
+2Single-language instructions
+3Description length 513: enough signal without eating the budget
+4Structure: 40 headings
+3Step-by-step instructions: 44 items
+3Output format is stated explicitly
+4Has examples (16 code blocks)
+1License stated

Quality base 70; lint remarks subtract, signals add up to 100. Result: 75.

External checks

ClawHub: suspicious

The skill mostly matches its trust-gate and verification purpose, but it needs review because some bundled examples undercut the stated security boundary for API keys and local webhook exposure.

LLM: suspicious (high) · 11 Sept 2026

Tests

— no evals.json

— no spec.yaml

— not run on models

Files · 43

.env.example 0K
LICENSE 1K
README.md 17K
SKILL.md 33K
javascript/README.md 4K
javascript/mcp/README.md 2K
javascript/mcp/mcp_test.mjs 7K
javascript/package-lock.json 29K
javascript/package.json 1K
javascript/quickstart/attest_action.mjs 2K
javascript/quickstart/extract_text.mjs 3K
javascript/quickstart/reconstruct_credential.mjs 2K
javascript/quickstart/test_integration.mjs 9K
javascript/quickstart/validate_document.mjs 2K
javascript/quickstart/verify_text.mjs 4K
javascript/use-cases/batch_verify.mjs 4K
javascript/use-cases/disinformation_shield.mjs 4K
javascript/use-cases/gate_tool_call.mjs 2K
javascript/use-cases/hallucination_audit.mjs 4K
javascript/use-cases/informational_credential.mjs 2K
javascript/use-cases/private_references.mjs 3K
javascript/use-cases/verify_article_url.mjs 3K
javascript/webhooks/webhook_receiver.mjs 3K
python/README.md 4K
python/mcp/README.md 4K
python/mcp/veritier_mcp_proxy.py 27K
python/mcp/veritier_mcp_test.py 11K
python/quickstart/attest_action.py 3K
python/quickstart/extract_text.py 3K
python/quickstart/reconstruct_credential.py 2K
python/quickstart/test_integration.py 9K
python/quickstart/validate_document.py 2K
python/quickstart/verify_text.py 4K
python/requirements.txt 0K
python/use-cases/batch_verify.py 4K
python/use-cases/disinformation_shield.py 4K
python/use-cases/gate_tool_call.py 2K
python/use-cases/hallucination_audit.py 3K
python/use-cases/informational_credential.py 2K
python/use-cases/private_references.py 3K
python/use-cases/verify_article_url.py 3K
python/webhooks/webhook_receiver.py 3K
skill-card.md 3K

When to use

**`attest_action`** is the Agent Trust Gate — the one that can stop the agent. Use it when a tool call depends on a record actually existing: an npm or PyPI package, a court opinion, an SEC filing, a section of the U.S. Code, or a policy corpus you supply. It returns a decision and a signed workpaper (Claim Credential) before the agent proceeds. A `block` decision means the call must not run. Verify never fail-closes; attest does. **`extract_text`** pulls the checkable claims out of a block of

Frontmatter

version: 3.4.2

homepage: https://veritier.ai

metadata: {"openclaw":{"requires":{"env":["VERITIER_API_KEY"]},"optional_env":{"VERITIER_TEST_KEY":"Optional test-environment key

Dates

In catalog since 11 Apr 2026

Updated in catalog 11 Sept 2026

First seen here 9 Sept 2026

Last analysis 12 Sept 2026, 23:07 UTC

Last seen in crawl 23 h ago

Score history

Date		Score	Safety	Qual.
12 Sept 2026	C	68	63	75
11 Sept 2026	C	68	63	75
9 Sept 2026	C	68	63	75