CLOUDM0N
Repository intelligence

sickn33/agentic-awesome-skills

GitHub

A local, agent-owned control plane and library for AI coding assistants to search, select, validate, and plan reproducible, structured skill stacks.

CLOUDM0N decision
Trust FAIL · 0/100
Good fit if

AI coding agents like Claude Code, Cursor, or Codex CLI that require a local, structured catalog of capabilities to perform specialized tasks

Watch out for

Does not verify or certify the semantic suitability, operational safety, or setup correctness of the selected skills

Practical intelligence

What matters before you adopt it

Problem it solves

The fleeting, context-exhausting nature of coding agent skill selections that are lost in chat histories, and the risk of assistant crash loops caused by overloading runtime directories with too many active skills.

Best for
AI coding agents like Claude Code, Cursor, or Codex CLI that require a local, structured catalog of capabilities to perform specialized tasks
Developers seeking to define, review, and persist immutable, reproducible plans of agent skill selections before altering a project surface
Main trade-offs
Does not verify or certify the semantic suitability, operational safety, or setup correctness of the selected skills
Composed stack manifests are limited to a technical maximum of 128 skills
The transactional apply and recovery features are explicitly experimental and remain disabled without active opt-in configuration
Why it stands out
Offers an extensive catalog of over 2,097 skills across development, security, testing, cloud infrastructure, and marketing
Generates dual output manifests (aas-stack.json and aas-selection-evidence.json) to record a factual process trace and capability ledger
Provides a browser-local, zero-install hosted Skill Workbench to visually inspect generated stacks and immutable plans in browser memory
Trust & CVEs

Security evidence without the noise

Trust remains a decision signal; CVEs and scanner evidence explain what is driving the risk.

Security findings
5
CLOUDM0N scanner findings
Critical
1
High
1
Medium
3
Low
0
View trust evidence & security findings▼
Why this score
CLOUDM0NCLOUDM0N scanner found a critical security issue.
TRIVYTrivy evidence is unavailable for this scan.
OPENSSFOpenSSF Scorecard evidence is unavailable for this scan.
CLOUDM0N findings
CRITICAL
Private-key material appears to be committed in executable/config scope. 1 sample match(es) found.
HIGH
Remote download piped or chained into a shell requires manual review. 5 sample match(es) found.
MEDIUM
Dynamic code execution pattern detected. 5 sample match(es) found.
MEDIUM
The project can spawn operating-system processes; review command construction and input handling. 5 sample match(es) found.
MEDIUM
Broad permission or elevated-command pattern detected. 5 sample match(es) found.
Architecture from code7 modules · 0 edges▼
Structural evidence

Modules and dependency edges extracted from repository code. This is code evidence, not README inference.

Code files
333
Modules
7
Dependency edges
0
Core modules
apps
56 files
plugins
276 files
Dependency flow
No dependency edges were extracted.
Detected languages
Python · TypeScript · Shell · JavaScript · Go · Java · Swift
Detected frameworks
Anthropic SDK · FastAPI · NumPy · Requests · pandas
Architecture evidence details
flowchart TD
    %% agentic-awesome-skills — high-level architecture (DRAFT, refine me)
    n0["apps · 56 files"]
    n1["docs · documentation · 1 file"]
    n2["plugins · 276 files"]
    class n1 docs
    classDef docs fill:#9d7660,color:#ffffff,stroke:#7c5d4c
Evidence, security & integrations▼
Integrations
Claude CodeCursorCodex CLIGemini CLIAutohand CodeAntigravity (IDE & CLI)Kiro (IDE & CLI)GitHub Copilot
Security notes
• All local MCP search, inspection, and validation operations run completely offline and never upload project metadata to external endpoints
• The installer validates package commits against the immutable gitHead hash recorded in npm metadata to prevent malicious tag-drift attacks
• Windows setups run PowerShell Get-Acl checks to enforce strict permission and access boundaries, raising 'AAS_ADAPTER_WINDOWS_ACL_FAILED' on validation errors
Still unknown
• The README does not specify recommended hardware requirements, minimum system RAM, or benchmark latency figures for running the local MCP server
• Does not detail how custom third-party integrations should authenticate or establish transport protocols with the aas-mcp daemon
• Lacks configuration instructions showing how the experimental 'apply' and 'recovery' execution engines are enabled or configured
Adoption guidance▼
Adopt if
+ You run advanced AI development agents (like Claude Code or Cursor) and need a reliable, local, and version-controlled way to track and pin agent behaviors
+ You want to enforce structured, evidence-based reporting (aas-selection-evidence.json) of what tools and capabilities are chosen during an agent session
+ You require an isolated, offline-first skill directory setup that protects intellectual property and handles sensitive workflows locally
Avoid if
– You expect an out-of-the-box framework that certifies the semantic correctness, runtime safety, or system compatibility of the tools chosen by the agent
– You require a fully validated, production-ready automated 'apply' and 'recovery' system, which currently remains highly experimental
– Your target assistant cannot handle selective activation and you are unable to apply context-pruning filters to manage large skill directories
How it works & getting started▼
How it works
1.A compatible AI agent uses the read-only AAS Core MCP server to search and inspect the local skill catalog
2.The agent evaluates the project surface and selects non-redundant skill IDs to meet the required capabilities
3.The read-only compose_stack tool validates the structural integrity and IDs of the selection in memory
4.The CLI or client persists the selection as aas-stack.json and a selection evidence sidecar file in the project's directory
5.The CLI command 'aas stack plan' processes the stack manifest to output an immutable, per-target plan without applying any live changes
Getting started
→ Configure the local MCP server by following the steps outlined in the AAS Core preview guide
→ To install a focused skill for GitHub Copilot, run: gh skill install sickn33/agentic-awesome-skills skills/brainstorming/SKILL.md --agent github-copilot --scope user --pin v14.2.0
→ To install skills for Cursor, execute: npx agentic-awesome-skills --cursor
→ To initialize a safe, narrow subset for OpenCode, use: npx agentic-awesome-skills --path .agents/skills --category development,backend --risk safe,none
Agent handoff
Use with any agent
JSON API
Alternatives

Nearby repositories worth comparing before adoption.

Compare top options →
codex-settings
feiskyer/codex-settings
49
Fit

Curated skills, profiles, and config templates that supercharge Codex CLI — multi-model, deep research, image gen & browser automation.

Trust REVIEW · 90
Compare →
awesome-claude-code
hesreallyhim/awesome-claude-code
63
Fit

A hand-picked collection of the finest of resources for the most awesome of agents, Claude Code, the undisputed champion of coding companions, from the unstoppable team at Anthropic PBC. A delectable showcase of top tier skills, ambidextrous agents, scintillating status lines, top notch developer tooling, and also we have plugins

Trust REVIEW · 90
Compare →
SztuCode
rojim666/SztuCode
34
Fit

本地优先的 AI 编程 Agent，支持 TUI/桌面端、工具权限、会话记忆、Skills、Subagents 与 MCPA local-first AI coding agent with TUI and desktop clients, tool permissions, memory, Skills, Subagents, and MCP support.。

Trust REVIEW · 45
Compare →