#1 · Development
GitHub Automation
High-utility workflow skill for repositories, issues, pull requests, and developer automation.
88
Overall
91
Trust
74
Momentum
Low riskRequires GitHub tokenNetwork accessStrong business utility
SR
SKILLRADAR
AI Security • Benchmarking • Index
Menu⌄
Safety Radar beta
Security and benchmarking for the tools your agents run.
SkillRadar is the evaluation layer for agent automation dependencies: security scanning today, performance benchmarks next, and private registries plus runtime skill recommendations for teams adopting agents.
Radar preview
Safety Radar report
Review signal
New shell command detected
A watched agent skill changed its install/run instructions. SkillRadar flags the command, source file, risk level, and date so a human can review before automation.
0
changes scanned
0
high-review signals
1443
candidates discovered
Four Radar pillars
Not a marketplace. The evaluation layer.
GitHub and agent hubs can host files. SkillRadar scores trust, safety, quality, compatibility, and performance — then exposes the signal to humans and agents.
beta now
Safety Radar
Automated security and clean-code scanning for prompt injection risk, hidden curl | bash commands, credential leaks, external domains, and risky diffs.
next
Performance Radar
Live benchmarks showing task success rate, speed, token efficiency, and model compatibility instead of vague skill claims.
enterprise
Private Radar
Company-owned private registries for vetted internal skills, approved versions, policy checks, and governance around public dependencies.
runtime API
AgentSkillOS
Semantic orchestration that tells agents which safe, high-performing skill stack to use for a complex task without instruction bloat.
Agent Skill Kits
Start with a watched bundle, not a random skill.
Curated bundles for common agent workflows. Each kit shows recommended skills, trust/risk mix, source links, and a watchlist path for future changes.
1 skills · avg trust 92
Builder Kit
Make a coding agent better at planning, API review, tests, and shareable diffs.
1 Low
3 skills · avg trust 78
Research Kit
Give your agent web research, community search, repo issue search, and durable notes.
1 Low / 2 Medium
4 skills · avg trust 90
Product Growth Kit
Help an agent improve landing pages, ecommerce funnels, sourcing, and experiments.
4 Low
Leaderboard preview
Top skills worth watching
#2 · Web Automation
Agent Browser
Useful browser automation layer for scraping, forms, screenshots, and web workflows.
84
Overall
77
Trust
89
Momentum
Medium riskBroad browser permissionsCan access sensitive sessionsReview before production use
#3 · Productivity
Obsidian
Practical personal knowledge-management integration with a clear local-file workflow.
81
Overall
86
Trust
62
Momentum
Low riskFilesystem accessLow external dependency surface
#4 · Productivity
Summarize
Straightforward document and URL summarization. Simple, useful, and easy to understand.
79
Overall
82
Trust
70
Momentum
Low riskMay send content to model APIsCheck data handling
#5 · Discovery
Find Skills
Meta-skill for finding other skills. Useful, but should not be allowed to install blindly.
76
Overall
73
Trust
77
Momentum
Medium riskCan influence installed toolingRequire human approval for installs
Latest changes
What changed since the last scan
0 total changes · 0 new skills · 0 high-review signals. These are automated review flags, not install recommendations.
Scan: 2026-06-21
No review-significant changes detected since the previous scan. The snapshot archive is still growing in the background.
Trust analysis sample
Review signals before you install
SkillRadar reviews public skill descriptions as text only, tracks changes over time, and surfaces permission/risk signals before you install. It does not execute third-party code, install dependencies, or load scanned skills. Each promoted report is moving toward trust-before-install provenance with source hashes and evidence snippets.
#1 · Trust analysis
obsidian-vault-maintainer
Maintain an Obsidian-friendly memory wiki vault with wikilinks, frontmatter, and official Obsidian CLI awareness.
88
Overall
92
Trust
95
Momentum
Low riskNo high-signal flags in text scan
#2 · Trust analysis
openclaw-landable-bug-sweep
Find or repair small high-confidence non-SDK-boundary OpenClaw bugfix PRs until five are landable.
88
Overall
84
Trust
95
Momentum
Low riskshell_command
#3 · Trust analysis
product-marketing-context
When the user wants to create or update their product marketing context document. Also use when the user mentions 'product context,' 'service context,' 'marketing context,' 'set up context,' 'positioning,' 'who is my target audience,' 'describe my product,' 'describe my service,'
85
Overall
92
Trust
95
Momentum
Low riskNo high-signal flags in text scan
#4 · Trust analysis
aes-cart-abandonment-analyzer
Identify reasons for cart abandonment and build multi-touch recovery sequences across email, SMS, and push.
85
Overall
92
Trust
95
Momentum
Low riskNo high-signal flags in text scan
Risk watch
Flags worth reviewing before install
High review
Free / Unlimited AI access skills
Positioning that promises unlimited paid AI access is a review trigger. Check outbound calls, credential requests, and hidden service dependencies.
Medium review
Auto-updater style skills
Useful if trustworthy, dangerous if not. Any skill that can modify installed skills deserves strict review and version-diff alerts.
High review
Skills with remote shell install commands
curl | bash, wget installers, and opaque package scripts should be flagged until manually inspected.
Get the signal
Join the SkillRadar launch list.
Daily source refreshes, top skills, risk flags, and early access to private scans, watchlists, and team allowlists.
Prefer the hosted beehiiv page? Open it here →