How protechnocrats/algorithmic-intelligence works
Algorithmic Intelligence is an agent skill for performance work: an agent loads its instructions, analyzes a codebase, proposes and implements an approved optimization, verifies behavior, then measures performance. Its knowledge base supplies technique and language-specific guidance. Optional plugins add decision guardrails and exact benchmark, growth-fit, and differential-test helpers. The README documents the workflow and plugin capabilities; sampled source confirms MCP calculations and hook behavior, but does not show all runtime wiring, so unsampled connections are omitted.
Main entry points: server.py MCP server · SKILL.md Skill instructions · references Technique references · languages Language guides
Components 20
Optimization workflow
- Skill instructions
skills/algorithmic-intelligence/SKILL.md - Understand problem
skills/algorithmic-intelligence/SKILL.md - Analyze solution
skills/algorithmic-intelligence/SKILL.md - Explore alternatives
skills/algorithmic-intelligence/SKILL.md - Engineer solution
skills/algorithmic-intelligence/SKILL.md - Verify correctness
skills/algorithmic-intelligence/SKILL.md - Measure performance
skills/algorithmic-intelligence/SKILL.md
Reference knowledge
- Technique references
skills/algorithmic-intelligence/references - Language guides
skills/algorithmic-intelligence/references/languages - Technique guides
skills/algorithmic-intelligence/references/techniques
Optional guard plugin
- Command guard
plugin/scripts/guard.py - Edit gate
plugin/scripts/edit_gate.py - Coverage hook
plugin/scripts/coverage_check.py - Summary hook
plugin/scripts/summary_check.py
MCP helpers
- MCP server
plugin/mcp/server.py - Benchmark comparison
plugin/mcp/server.py - Growth fit
plugin/mcp/server.py - Differential checks
plugin/mcp/server.py
Other components
- Coding agent
- Reference checker
tools/check_refs.py
Connections 19
- Coding agent → Skill instructions: loads
README.md - Skill instructions → Understand problem: guides
README.md - Understand problem → Analyze solution: then analyze
README.md - Analyze solution → Explore alternatives: then explore
README.md - Explore alternatives → Engineer solution: then engineer
README.md - Engineer solution → Verify correctness: then verify
README.md - Verify correctness → Measure performance: then measure
README.md - Skill instructions → Technique references: loads as needed
README.md - Technique references → Language guides: includes
README.md - Technique references → Technique guides: includes
README.md - Coding agent → Command guard: requests guarded actions
README.md - Coding agent → Edit gate: requests code edits
plugin/scripts/edit_gate.py - Coding agent → Coverage hook: ends session
plugin/scripts/coverage_check.py - Coding agent → Summary hook: submits summary
plugin/scripts/summary_check.py - Coding agent → MCP server: calls tools
README.md - MCP server → Benchmark comparison: dispatches
plugin/mcp/server.py - MCP server → Growth fit: dispatches
plugin/mcp/server.py - MCP server → Differential checks: dispatches
plugin/mcp/server.py - Coding agent → Reference checker: checks references