panther-intent-capture
ExtensionLocal-first browser intent capture with visible controls, evidence-linked proposals, portable JSON, and opt-in verified ticket delivery.
Mike Hamanaka ยท Open-source tools
Local tools. Visible evidence. Human control.
Explore tools for capturing intent, coordinating AI coding sessions, and verifying the work.
Explore a project's source or read its documentation. Each repository describes its supported workflows, setup, license, and current status.
Local-first browser intent capture with visible controls, evidence-linked proposals, portable JSON, and opt-in verified ticket delivery.
Ticket standard for local-first, evidence-driven coordination with bounded context and visible failures.
Local-first coordination core for AI CLI sessions with bounded context, explicit ownership, and evidence trails.
Hardware Air Traffic Control for autonomous AI mesh - thermal/power/memory monitoring and guardrailed dispatch.
Open Edge Topology mapping - empirical infrastructure discovery, diffing, and self-documentation.
Autonomous gatekeeper for independent intent verification, guardrailed retry loops, and proof-of-work commits.
Local-first memory core - extracts failure heuristics via local Ollama and commits hardened rules.
Asynchronous inference queue and fault-tolerance manager for heterogeneous edge fleet.
These projects explore how local-first tools can turn intent into reviewable work. Each component has a focused role; permissions, execution, and acceptance remain explicit decisions.
The sequence below is a conceptual workflow, not a deployment or integration status report. Use the project directory above to find each public repository.
An illustration of the intended workflow and verification feedback. Consult each repository for implemented behavior.
Read the workflow as: capture โ distill โ triage โ sandbox โ coordinate hardware โ execute โ verify โ record. Proposed retries return to triage for review.