Skip to content

Commit 5a6b44a

Browse files
Nick Sullivanclaude
Nick Sullivan
and
claude
committed
📊 Add competitor analysis for 6 AI coding config projects
Researched and documented top competitors by GitHub stars: - awesome-cursorrules (36k stars) - Cursor rules collection - awesome-claude-code (18k stars) - Claude Code resources - superpowers (10k stars) - Claude Code skills library - Personal_AI_Infrastructure (1.4k stars) - Full AI OS by Daniel Miessler - dotai (1.1k stars) - Cross-platform context manager - ai-prompts (857 stars) - Multi-tool prompt collection Repos cloned to ../reference/ for detailed analysis. Knowledge files follow prompt-engineering best practices for LLM-readable content. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <[email protected]>
1 parent cfffa67 commit 5a6b44a

File tree

6 files changed

+296
-0
lines changed

6 files changed

+296
-0
lines changed

‎knowledge/competitors/ai-prompts.md‎

Lines changed: 38 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,38 @@
1+
# instructa/ai-prompts
2+
3+
Multi-tool prompt collection. 857 stars. Curated prompts for Cursor, Cline, Windsurf,
4+
and GitHub Copilot.
5+
6+
GitHub: https://github.com/instructa/ai-prompts
7+
Local: ../reference/ai-prompts
8+
9+
## What They Do
10+
11+
Open-source prompt collection for multiple AI coding assistants. Focus on project
12+
scaffolding and coding standards. Part of the larger Instructa AI tooling ecosystem.
13+
14+
## Why They Matter
15+
16+
Multi-tool support without locking to one assistant. Well-organized prompt library with
17+
good categorization. Practical focus on scaffolding and standards.
18+
19+
## What We Can Learn
20+
21+
Prompt organization: Category structure. Naming conventions. Directory layout.
22+
23+
Multi-tool handling: Tool-specific format differences. Common patterns across tools.
24+
Translation approaches between formats.
25+
26+
Content types: Popular prompts. Scaffolding patterns. Coding standards templates.
27+
28+
## Key Files to Review
29+
30+
Directory structure for prompt organization. Tool-specific formatting approaches.
31+
README for documentation style.
32+
33+
## Our Differentiation
34+
35+
Plugin system instead of just files. Agents and skills beyond prompts. Philosophy layer
36+
with values. Bootstrap automation instead of manual setup.
37+
38+
They're lighter-weight but less comprehensive than our approach.

‎knowledge/competitors/awesome-claude-code.md‎

Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
# hesreallyhim/awesome-claude-code
2+
3+
Claude Code resources collection. 18,238 stars. Curated list of slash-commands,
4+
CLAUDE.md files, CLI tools, workflows, and guides.
5+
6+
GitHub: https://github.com/hesreallyhim/awesome-claude-code
7+
Local: ../reference/awesome-claude-code
8+
9+
## What They Do
10+
11+
Comprehensive Claude Code resource list. Goes beyond rules to include commands,
12+
workflows, tooling, and patterns. Community-driven with active contributions.
13+
14+
## Why They Matter
15+
16+
Direct competitor in our exact space. Fast growth indicates strong demand for Claude
17+
Code resources. Shows what types of content the community wants most.
18+
19+
## What We Can Learn
20+
21+
Content categories: What resources developers seek. How to organize Claude Code
22+
materials. Popular command and workflow patterns.
23+
24+
Presentation: README structure. Information hierarchy. Category organization.
25+
26+
## Key Files to Review
27+
28+
README.md for organization approach. Command examples for popular patterns. CLAUDE.md
29+
templates for configuration approaches.
30+
31+
## Our Differentiation
32+
33+
Plugin marketplace instead of just a list. Install command integration. Agent and skill
34+
definitions. Structured philosophy and values baked in.

‎knowledge/competitors/awesome-cursorrules.md‎

Lines changed: 42 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,42 @@
1+
# PatrickJS/awesome-cursorrules
2+
3+
Cursor AI rules collection. 36,249 stars. The definitive community collection of
4+
.cursorrules files organized by language, framework, and use case.
5+
6+
GitHub: https://github.com/PatrickJS/awesome-cursorrules
7+
Local: ../reference/awesome-cursorrules
8+
9+
## What They Do
10+
11+
Community-curated .cursorrules files for Cursor AI. Ready-to-use configurations
12+
organized by category. Simple model with no tooling - just copy files.
13+
14+
## Why They Matter
15+
16+
Massive community traction makes this the clear space leader. Hundreds of battle-tested
17+
rules covering Angular, React, Vue, Python, Go, Rust, and more. The content itself is
18+
a goldmine for understanding what users want from AI coding configurations.
19+
20+
## What We Can Learn
21+
22+
Content patterns: How they organize rules by language, framework, use case. What
23+
patterns appear across high-quality .cursorrules files.
24+
25+
Community model: PR-based contributions. How they maintain quality. How they became the
26+
go-to resource for Cursor users.
27+
28+
## Limitations (Our Opportunity)
29+
30+
Static files only with no plugin system. Cursor-specific so it doesn't work with Claude
31+
Code. No versioning or update mechanism. Manual copy-paste installation.
32+
33+
## Key Files to Review
34+
35+
README.md for organization and presentation. The rules/ directory for actual content.
36+
Contributing guidelines for community engagement model.
37+
38+
## Our Differentiation
39+
40+
Plugin-based distribution instead of copy-paste. Cross-tool support for both Cursor and
41+
Claude Code. Commands, agents, and skills beyond just rules. Heart-centered philosophy
42+
layer.

‎knowledge/competitors/dotai.md‎

Lines changed: 43 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,43 @@
1+
# udecode/dotai
2+
3+
Cross-platform AI context manager. 1,107 stars. Generates configurations for multiple
4+
AI tools from a single source using the skiller CLI.
5+
6+
GitHub: https://github.com/udecode/dotai
7+
Local: ../reference/dotai
8+
9+
## What They Do
10+
11+
Context manager for all coding agents. Single source format generates CLAUDE.md,
12+
AGENTS.md, Cursor rules, and MCP configurations. Uses `npx skiller@latest apply` for
13+
generation.
14+
15+
## Why They Matter
16+
17+
Closest conceptual competitor with similar cross-platform vision. Active development
18+
with regular updates. Generation approach is an alternative to our symlink strategy.
19+
20+
## What We Can Learn
21+
22+
Architecture: Source format structure. Generation pipeline for multiple outputs. Rule
23+
organization and inheritance patterns.
24+
25+
CLI design: Installation experience. Update mechanisms. Configuration management.
26+
27+
Content: Rules and skills included. Session memory patterns. PR workflow automation.
28+
29+
## Key Files to Review
30+
31+
Source rule format for input specification. Generation scripts for output handling.
32+
Skill definitions for behavior patterns. README for positioning and UX.
33+
34+
## Key Differences
35+
36+
We use symlinks, they use generation. We're plugin-marketplace-first, they're CLI-first.
37+
We have bootstrap.sh, they have skiller CLI. Our philosophy is more explicit with
38+
heart-centered values.
39+
40+
## Questions to Explore
41+
42+
Could we learn from their generation approach? Is symlink-based simpler or more complex
43+
for users? What content do they have that we should add?

‎knowledge/competitors/personal-ai-infrastructure.md‎

Lines changed: 102 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,102 @@
1+
# danielmiessler/Personal_AI_Infrastructure
2+
3+
Complete personal AI operating system. 1,442 stars. Open-source scaffolding for
4+
building your own AI-powered system with full philosophy, architecture, and operational
5+
framework.
6+
7+
GitHub: https://github.com/danielmiessler/Personal_AI_Infrastructure
8+
Local: ../reference/Personal_AI_Infrastructure
9+
Author: Daniel Miessler (creator of fabric, 30k+ stars)
10+
11+
## What They Do
12+
13+
Full AI operating system, not just configs. Includes 1500-line Constitution with 13
14+
founding principles. Skills, agents, hooks, history capture, voice feedback, MCP
15+
management. Tagline: "Personal AI Infrastructure for upgrading humans."
16+
17+
## Why They Matter
18+
19+
Most sophisticated competitor in the space. Daniel brings production experience from
20+
building fabric. Complete system with strong opinions on CLI-First and Code Before
21+
Prompts. Deep philosophical foundation.
22+
23+
## Architecture
24+
25+
13 Founding Principles: Clear Thinking is King. Scaffolding > Model. As Deterministic
26+
as Possible. Code Before Prompts. Spec/Test/Evals First. UNIX Philosophy. ENG/SRE
27+
Principles. CLI as Interface. Goal > Code > CLI > Prompts > Agents. Meta/Self Update.
28+
Custom Skill Management. Custom History. Custom Agent Voices.
29+
30+
Four Primitives: Skills (domain expertise containers), Commands (task workflows within
31+
skills), Agents (autonomous executors with 12+ distinct voices), MCPs (external tool
32+
integrations).
33+
34+
Three-Tier Context Loading: Tier 1 is system prompt (always active, 200-500 words).
35+
Tier 2 is SKILL.md body (on-demand, 500-2000 lines). Tier 3 is reference files
36+
(just-in-time deep dives).
37+
38+
Key Systems: Voice Server with ElevenLabs TTS for spoken feedback. History System for
39+
automatic work capture. Hook System for event-driven automation. MCP Profiles for
40+
context-specific tool configurations.
41+
42+
## What We Can Learn
43+
44+
Philosophy: "Build the scaffolding first, then add the AI." "Code is cheaper, faster,
45+
and more reliable than prompts." "Build tools that work perfectly without AI, then add
46+
AI to make them easier to use."
47+
48+
CLI-First Architecture: Every operation accessible via CLI. Code before prompts.
49+
Deterministic over probabilistic.
50+
51+
Skills-as-Containers: Not just markdown docs but active orchestrators with routing,
52+
workflows, and assets.
53+
54+
Progressive Disclosure: Token-efficient context loading. Only load what's needed.
55+
56+
Two-Tier MCP Strategy: Tier 1 for discovery (exploratory, high token cost). Tier 2 for
57+
production (TypeScript wrappers, efficient).
58+
59+
Operational Patterns: Structured output format with COMPLETED line for voice. Spotcheck
60+
pattern for verifying parallel work. Scratchpad vs History separation.
61+
62+
## Key Files to Review
63+
64+
CONSTITUTION.md for 1500 lines of architectural philosophy. .claude/skills/CORE/SKILL.md
65+
for main system skill. .claude/skills/CORE/SkillSystem.md for skill creation guide.
66+
.claude/hooks/ for event automation. .claude/agents/ for 12+ agent definitions.
67+
68+
## Differentiation Analysis
69+
70+
What they have that we don't: Voice integration with ElevenLabs TTS. History capture
71+
system. MCP profile management. Observability dashboard. More opinionated CLI-First
72+
stance.
73+
74+
What we have that they don't: Cross-tool support for Cursor and Claude Code together.
75+
Plugin marketplace distribution. Heart-centered philosophy. Simpler installation with
76+
bootstrap.sh.
77+
78+
Philosophical differences: PAI focuses on "Scaffolding > Model" with deterministic
79+
systems. We focus on "Heart-centered engineering" with human flourishing. Both valid
80+
but different orientations. PAI is engineering-first, we're values-first.
81+
82+
## Ideas to Consider
83+
84+
Borrow: Progressive disclosure pattern with 3-tier loading. Structured output format
85+
concept. Scratchpad vs History separation.
86+
87+
Consider: Voice feedback integration. More explicit principles or constitution.
88+
89+
Keep: Our simpler plugin-based distribution model.
90+
91+
## Competitive Positioning
92+
93+
PAI is comprehensive but complex. Our opportunity is simpler entry point (install a
94+
plugin vs clone entire system), cross-tool support (Cursor users can use us while PAI
95+
is Claude Code only), philosophy layer (heart-centered vs engineering-first), and
96+
modular approach (pick what you want vs all-or-nothing).
97+
98+
## Notes
99+
100+
Daniel is building the "AI operating system for individuals" with same vision but
101+
different execution. Worth watching as both competitor and inspiration. The Constitution
102+
alone is worth reading in full as a textbook on AI system architecture.

‎knowledge/competitors/superpowers.md‎

Lines changed: 37 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,37 @@
1+
# obra/superpowers
2+
3+
Claude Code skills library. 10,148 stars. Core skills that enhance Claude Code with
4+
specialized behaviors and workflows.
5+
6+
GitHub: https://github.com/obra/superpowers
7+
Local: ../reference/superpowers
8+
9+
## What They Do
10+
11+
Pre-built skills library for Claude Code. Focus on reusable behavior patterns. Works
12+
with Claude Code's official plugin system.
13+
14+
## Why They Matter
15+
16+
Jesse Vincent (obra) is a well-known developer with quality standards. 10k stars in
17+
approximately 2 months since plugins launched indicates strong demand. Skills-focused
18+
approach complements our broader scope.
19+
20+
## What We Can Learn
21+
22+
Skill design: How they structure definitions. What distinguishes a skill from a command
23+
or agent. Composition patterns.
24+
25+
Implementation quality: Code organization. Testing approaches. Documentation standards.
26+
27+
Distribution: Installation handling. Versioning and updates. User onboarding experience.
28+
29+
## Key Files to Review
30+
31+
Skill definitions for structure and patterns. Plugin manifest for distribution approach.
32+
Documentation for user experience.
33+
34+
## Our Differentiation
35+
36+
Broader scope beyond just skills. Cross-tool support for Cursor and Claude Code.
37+
Philosophy and values integrated. Bootstrap script for full setup.

0 commit comments

Comments
 (0)