🔎 Evidence browser

Browse the trust index

Search by skill, publisher, category, or trust summary — then use the runtime filters to find cards with live test evidence. The two main lanes are baseline safety checks first and deeper follow-on functionality checks after that.

⚙️ Filters · 1 active
✨ Quick picks
🏷 Categories · coding-agents-and-ides

🧾 Evidence level: source-scanned means local source evidence; catalog-only means thinner metadata-first coverage.

🧪 Runtime status: cards can show only the baseline safety lane or the deeper follow-on functionality lane, depending on how far the skill got. Some cards now also surface how the skill behaved when clearly fake credentials were present.

📏 Depth cue: tells you whether the evidence stops at baseline checks, includes follow-on functionality checks, or includes richer fixture/example proof.

⏱ Freshness cue: tells you whether the latest runtime evidence is from the last 24 hours, the last 7 days, or is older and therefore less current.

🩺 Failure confidence: distinguishes a first seen failure from a repeated failure or a regression after an earlier pass, so not every red row means the same thing.

🧪 Fake-auth behavior: when available, this tells you whether a skill handled clearly fake credentials cleanly, needed real access to continue, or behaved badly around credential-like input.

Results

Showing 24 of 1200 skills in the browsable catalog view · category: coding-agents-and-ides · sort: score
This snapshot is for the current page of results, not the whole filtered universe.
Browse hint: slices with zero failures plus some source-scanned or reviewed entries deserve more attention first; fresh runtime evidence helps too, because old clean receipts can still hide current drift.

internal-linking-optimizer

aaron-he-zhu · vsource-scanned
54
overall

This skill should be used when the user asks to "fix internal links", "improve site architecture", "link structure", "distribute page authority", "internal linking strategy", "orphan pages", "site architecture is messy", or "pages have no links pointing to them". Analyzes and optimizes internal link structure to improve site architecture, distribute page authority, and help search engines understand content relationships. Creates strategic internal linking plans. For a broader on-page audit, see on-page-seo-auditor. For external link analysis, see backlink-analyzer.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (trading, email), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

iterative-code-evolution

aaronjmars · vsource-scanned
54
overall

Systematically improve code through structured analysis-mutation-evaluation loops. Adapted from ALMA (Automated meta-Learning of Memory designs for Agentic systems). Use when iterating on code quality, optimizing implementations, debugging persistent issues, or evolving a design through multiple improvement cycles. Replaces ad-hoc "try and fix" with disciplined reflection, variant tracking, and principled selection of what to change next.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found normal operational surface via environment, network, or shell-related references.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

kagi-summarizer

joelazar · vsource-scanned
54
overall

Summarize any URL or text using Kagi's Universal Summarizer API. Supports multiple engines (including the enterprise-grade Muriel model), bullet-point takeaways, and output translation to 28 languages. Use when you need a high-quality summary of an article, paper, video transcript, or any document.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

memos

fty4 · vsource-scanned
54
overall

This skill provides a simple interface to the Memos API (https://usememos.com/docs/api). It lets you create, read, delete, and list memos from OpenClaw.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

mh-things-mac

mohdalhashemi98-hue · vsource-scanned
54
overall

Manage Things 3 via the `things` CLI on macOS (add/update projects+todos via URL scheme; read/search/list from the local Things database). Use when a user asks OpenClaw to add a task to Things, list inbox/today/upcoming, search tasks, or inspect projects/areas/tags.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

quack

jpaulgrayson · vsource-scanned
54
overall

Agent-to-agent messaging, identity, and coordination via the Quack Network. Use when sending messages to other AI agents, checking your agent inbox, registering on the Quack Network, participating in challenges, or coordinating work with other agents. Triggers on "send a message to another agent", "check my quack inbox", "register on quack", "agent challenge", "inter-agent communication", "QuackGram", or "QUCK tokens".

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (wallet, token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

quantumos

murtiurti4 · vsource-scanned
54
overall

Install and manage QuantumOS, an AI command center dashboard for OpenClaw. Use when the user wants to set up QuantumOS, start/stop the dashboard, troubleshoot it, or update it. Provides a real-time chat UI, Mission Control (kanban task management with AI agents), and a feed dashboard (Reddit, HN, X, news).

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

quick-reminders

lstpsche · vsource-scanned
54
overall

Zero-LLM one-shot reminders (<48h) via nohup sleep + openclaw message send, operated via {baseDir}/scripts/nohup-reminder.sh.

Use Cautionconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token, telegram, whatsapp), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

requesting-code-review

zlc000190 · vsource-scanned
54
overall

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found normal operational surface via environment, network, or shell-related references.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

securevibes-scanner

anshumanbh · vsource-scanned
54
overall

Run AI-powered application security scans on codebases. Use when asked to scan code for security vulnerabilities, generate threat models, review code for security issues, run incremental security scans, or set up continuous security monitoring via cron. Supports full scans (one-shot) and incremental scans (cron-driven, only new commits).

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token, oauth), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

signal-messenger-standalone

lucksus · vsource-scanned
54
overall

Full Signal messenger integration for OpenClaw agents. Send/receive text and voice messages via signal-cli with role-based permissions (owner/trusted/untrusted), automatic voice transcription (Whisper), voice replies (TTS), conversation history, new contact triage, typing indicators, read receipts, and instant wake-on-message via OpenClaw hooks API. Use when setting up Signal messaging, handling Signal contacts, sending/receiving Signal messages, or managing Signal contact permissions.

Use Cautionconfidence: source evidencesource-scanned
+ 1 more
suspicious
Take: Potentially suspicious implementation signals detected: sudo .
Decision cue: Proceed carefully — suspicious signals matter more than capability surface alone.

skill-hub

phenixstar · vsource-scanned
54
overall

OpenClaw skill discovery, security vetting & install. Searches 3000+ curated skills from ClawHub registry and awesome-openclaw-skills catalog. Scores credibility, detects prompt injection & malicious patterns, manages installations. Quick-checks GitHub for new skills.

High Riskconfidence: source evidencesource-scanned
+ 1 more
suspicious
Take: Potentially suspicious implementation signals detected: eval(, password.
Decision cue: Proceed carefully — suspicious signals matter more than capability surface alone.

solo-swarm

fortunto2 · vsource-scanned
54
overall

Launch 3 parallel research agents (market, users, tech) to investigate an idea from multiple angles simultaneously. Use when user says "swarm research", "parallel research", "investigate fast", "3 agents", "team research", or wants faster alternative to /research. Produces research.md. Do NOT use for solo research (use /research) or idea scoring (use /validate).

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found normal operational surface via environment, network, or shell-related references.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

sovereign-quest-master

ryudi84 · vsource-scanned
54
overall

You are the Quest Master. You create daily quests, boss fights, and side quests tailored to the player's goals and current stats.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found normal operational surface via environment, network, or shell-related references.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

strykr-prism

nextfrontierbuilds · vsource-scanned
54
overall

Real-time financial data API for AI agents. Stocks, crypto, forex, ETFs. 120+ endpoints. Alternative to Alpha Vantage, CoinGecko. Works with Claude, Cursor.

Use Cautionconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (wallet, trading, token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

task-router-skill

capt-marbles · vsource-scanned
54
overall

Distributed task queue and agent coordinator for OpenClaw multi-agent systems. Route tasks to specialized agents by capability matching, track task lifecycle, handle async handoffs, rebalance loads, and manage dead letters. Use when: (1) Creating tasks programmatically or from heartbeats, (2) Routing work to specialized agents based on capabilities, (3) Monitoring task status and completion, (4) Coordinating multi-step workflows across agents, (5) Handling async agent work without blocking main sessions.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found normal operational surface via environment, network, or shell-related references.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

usememos

sinhong2011 · vsource-scanned
54
overall

>

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

usewhisper-autohook

alinxus · vsource-scanned
54
overall

Auto-hook tools for OpenClaw: query Whisper Context before every generation, ingest after every turn. Built for Telegram agents (stable user_id/session_id).

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token, telegram), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

agenticflow-skill

seanphan · vsource-scanned
53
overall

Comprehensive guide for building AI workflows, agents, and workforce systems with AgenticFlow. Use when designing workflows with various node types, configuring single agents, or orchestrating workforce collaboration patterns.

High Riskfollow-on functionality checks passed · 5/5confidence: source evidence
+ 2 more
source-scannedsuspicious
Runtime receipts + what passed2026-03-15 22:45 UTC
functionality-v2evidence depth: follow-on functionality checkstested recently: within 7 dayspassedoutput 80 Bartifacts 0worker oc-sandboxsource stage: cache hitsuite 1585 msbaseline-v3 8/8
RatioDaemon muttered: agenticflow-skill behaved itself under runtime pressure.5/5 functionality-v2 checks passed. Pleasantly boring.
Observed: skill-structure-ok
Take: Potentially suspicious implementation signals detected: password.
Decision cue: Proceed carefully — suspicious signals matter more than capability surface alone.

ai-agent-tools

cerbug45 · vsource-scanned
53
overall

This library provides ready-to-use Python functions that AI agents can leverage to perform various tasks including file operations, text analysis, data transformation, memory management, and validation.

Use Cautionconfidence: source evidencesource-scanned
+ 1 more
suspicious
Take: Potentially suspicious implementation signals detected: sudo .
Decision cue: Proceed carefully — suspicious signals matter more than capability surface alone.

ai-workforce

km2411 · vsource-scanned
53
overall

Turn an OpenClaw agent into an autonomous AI Chief that runs a business. Provides trust-based autonomy, structured knowledge management (bank/), worker delegation patterns, and reflection cycles. Use when setting up a new agent as a business operator, when onboarding a human, when delegating to sub-agents, when managing trust levels, or when running daily/weekly/monthly reflection and memory maintenance.

High Riskconfidence: source evidencesource-scanned
+ 1 more
suspicious
Take: Potentially suspicious implementation signals detected: password.
Decision cue: Proceed carefully — suspicious signals matter more than capability surface alone.

architecture-rendering

eftalyurtseven · vsource-scanned
53
overall

Generate photorealistic architectural renders and visualizations using each::sense AI. Create exterior views, interior renders, sketch-to-render conversions, and more for architects, designers, and real estate professionals.

Insufficient Evidenceconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found normal operational surface via environment, network, or shell-related references.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.

cacheforge-vibe-check

tkuehnl · vsource-scanned
53
overall

Audit code for "vibe coding sins" — patterns that indicate AI-generated code was accepted without proper review. Produces a scored report card with fix suggestions.

High Riskfollow-on functionality checks passed · 7/7confidence: source evidence
+ 2 more
source-scannedsuspicious
Runtime receipts + what passed2026-03-14 11:15 UTC
functionality-v2evidence depth: follow-on functionality checkstested recently: within 7 dayspassedoutput 117 Bartifacts 0worker oc-sandboxsource stage: cache hitsuite 2322 msbaseline-v3 8/8
RatioDaemon on this skillCacheforge Vibe Check is trying to handle cacheforge vibe check. Follow-on functionality checks currently pass without failed checks, the trust label is High Risk, and setup looks advanced.
Observed: skill-structure-ok
Take: Potentially suspicious implementation signals detected: eval(, rm -rf, password.
Decision cue: Proceed carefully — suspicious signals matter more than capability surface alone.

cc-godmode

cubetribe · vsource-scanned
53
overall

Self-orchestrating multi-agent development workflows. You say WHAT, the AI decides HOW.

Use Cautionconfidence: source evidencesource-scanned
+ 1 more
privileged capability
Take: Source-aware scan found higher-privilege capability areas (token, oauth, telegram, whatsapp, email), but that alone is not evidence of malicious behavior.
Decision cue: Decent evidence base — source-level signals are available, so inspect the receipts.