agent-loop-ext
Crash-resilient external agent loop with state persistence and CI/CD integration
Build a safe strict-LCM or adaptive polyrhythmic FlowGraph for explicit multi-track reasoning.
$ npx -y skills add jmagly/aiwg --skill polyrhythmic-reasoning --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/polyrhythmic-reasoningContext preview
The summary Claude sees to decide when to auto-load this skill.
Build a safe strict-LCM or adaptive polyrhythmic FlowGraph for explicit multi-track reasoning.
namespace: aiwg name: polyrhythmic-reasoning platforms: [all] description: Build a safe strict-LCM or adaptive polyrhythmic FlowGraph for explicit multi-track reasoning. triggers: - use polyrhythmic reasoning - build a 4 5 reasoning graph - compare strict lcm and adaptive reasoning - compose problem mode and user mode tracks
Use this skill for a non-high-risk task that benefits from explicit problem-mode and user-mode tracks.
1. Choose **strict-lcm** only when exact 4/5 alignment at activation 20 is the intended behavior. Choose **adaptive** when evidence-backed early convergence is allowed under the hard ceiling. 2. Use one of the approved domains: technical troubleshooting, conceptual explanation, practical planning, or theoretical comparison. 3. Keep user state **unknown** unless a user-stated evidence item supports a declared non-sensitive state. 4. Choose agent-only or agent-plus-read-only-tool. Never widen the latter past **filesystem:read** and **sideEffectMode: none**. 5. Build through **lib/polyrhythmic-reasoning.mjs**, validate the returned FlowGraph, and execute through the normal composition adapter. 6. Preserve conflicts and failed/skipped beats in typed state and events. Publish only the final synthesis.
Do not use the generic pattern for healthcare, financial advice, or another high-risk domain. Do not claim that the topology improves quality or efficiency without reproducible benchmark evidence.
See **docs/polyrhythmic-reasoning.md** and the four files under **examples/**.
Reusable project context and specialist workflows for the AI tools you already use. Plan software, coordinate specialist reviews, prepare campaigns, investigate incidents, organize research, curate media, and maintain operational knowledge.
Repo: jmagly/aiwg
Crash-resilient external agent loop with state persistence and CI/CD integration
Detect requests for iterative autonomous agent loops and route to the appropriate loop executor
Automatically execute tests when code-generating agents modify source files, enforcing the execute-before-return pattern
Enable agent loops to learn from similar past tasks and share patterns across loops
Query and manage the executable feedback debug memory
Execute tests on generated code and iterate until passing