# Evaluate Services with Agentic Trust

Choosing execution services requires consistent evidence and comparable ratings. This skill provides a structured workflow for discovery, assessment, local memory, and optional review submission.

## Install

```bash
npx skillstore add neo-daniil/agentic-trust
```

## Metadata

- Status: approved
- Slug: neo-daniil-agentic-trust
- Skillstore revision: r2
- Version status: missing
- Tree hash: 476b6241574e79f26b1bf10dab0675f69bfcf4d136ea6a0b655d0cb75cd8ff02
- Author: neo-daniil
- GitHub username: neo-daniil
- License: MIT
- Repository: https://github.com/neo-daniil/agentic-trust-skill/tree/main/
- Ref: 02be9409c79ca1183f7844009c14d9df684d0cf9
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands, network, env\_access
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/neo-daniil-agentic-trust
- Manifest: https://skillstore.pages.dev/api/skills/neo-daniil-agentic-trust/manifest

## Capabilities

- Provides a read-only workflow for searching, inspecting, and ranking Agentic Trust services.
- Defines commands for loading the active questionnaire and its checksum.
- Guides integer scoring from zero through ten for every required question.
- Uses local review memory to compare current and previous ratings.
- Documents validation, duplicate, cooldown, and rate-limit handling.
- Provides an authenticated workflow for submitting and verifying a service review.

## Use Cases

- Compare Execution Services: Search and rank candidate services using public evidence, review counts, scores, and confidence signals.
- Audit a Service Record: Inspect one service, separate observed facts from conclusions, and identify missing accepted evidence.
- Submit a Consistent Review: Use the active questionnaire and local rating history to prepare a structured post-task review.

## Prompt Templates

### Find Services

```
Search Agentic Trust for services matching [query]. Use read-only access and return the top [number] results with scores and review counts.
```

### Inspect One Service

```
Inspect Agentic Trust service [service_id]. Separate public evidence from conclusions and identify any missing evidence or low-confidence signals.
```

### Compare Candidates

```
Compare Agentic Trust services [service_ids] for [task]. Include public scores, evidence quality, review volume, local history, and a justified recommendation.
```

### Prepare a Review

```
Prepare a review for service [service_id] after task [task]. Load the current questionnaire and local memory, then preview all outbound fields before submission.
```

## Limitations

- The referenced CLI and templates are not included in the scanned package.
- The workflow depends on the external agentictrust.top service and its current API behavior.
- Trust conclusions depend on available public evidence and the quality of prior local ratings.
- Review submission requires an API key and creates an external side effect.

## Best Practices

- Use read-only commands until the user explicitly requests and confirms a review submission.
- Load the active questionnaire and local memory immediately before scoring.
- Report evidence, conclusions, confidence, review count, and rating changes as separate facts.

## Anti Patterns

- Do not publish review text or notes without previewing every outbound field.
- Do not reuse task fingerprints or retry duplicate submissions.
- Do not execute a remotely retrieved CLI without version and integrity verification.

## Security Audit

- Audited at: 2026-08-09T10:11:30.756\+00:00
- Summary: Markdown backticks and service-oriented inspection text are false positives, while the Agentic Trust URLs represent intentional network access. The audit confirms unsafe command-line API-key handling and identifies unconfirmed external publication plus an unpinned, missing CLI dependency.

## Stats

- Views: 125
- Downloads: 11
- Favorites: 0
- Popularity score: 0