Does Pathmode need an account or an API key?
No. Local mode is the default: specs live in `intent.md` and nothing leaves the machine. The preflight makes no model call.

@pathmodeio
Basic information
Deterministic preflight for the intent you hand to a coding agent: six calibrated gates, the exact blockers named, no model call, no key needed.
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"pathmode": {
"command": "npx",
"args": [
"@pathmode/mcp-server"
]
}
}
}Get the currently active intent, with a summary of the evidence it cites.
List intents with their status and objectives.
Get one intent by ID with its objective, outcomes, constraints, and edge cases.
Search intents by keyword across goals, objectives, outcomes, and constraints.
Run the deterministic six-check preflight on an intent and name the exact blockers. No model call.
Save an intent spec to the repository's intent.md, and to the workspace when one is connected.
Adopt an existing OpenSpec change as intent.md without modifying the change folder.
Confirm a readiness check the preflight read but could not confirm, or waive it with a person's approval.
Record a person's explicit answer to one open product choice in the local intent.md.
Export an intent as .cursorrules, a CLAUDE.md or AGENTS.md section, an Outcomes rubric, or an OpenSpec change.
Pathmode finds the product decisions a request leaves open, before your coding agent builds. It gives your agent check_intent_readiness, a deterministic preflight: six calibrated checks, no model call, the same verdict every run. Specs live in intent.md in your repository, as plain markdown you own, in the open IntentSpec format.
Add npx @pathmode/mcp-server to your MCP client config. No signup and no API key. Then ask your agent for a change the way you normally would. The server instructs it to run preflight first and, when the repository has no intent.md, to draft one, record the open product choices, and stop before changing code. These are instructions, not enforcement: whether an agent follows them is up to the agent. Optionally run npx @pathmode/mcp-server install-skills to add 9 Claude Code skills.
.cursorrules, CLAUDE.md, or AGENTS.md; import an existing OpenSpec changeintent.md for a change, with open choices recorded instead of guessedNo. Local mode is the default: specs live in `intent.md` and nothing leaves the machine. The preflight makes no model call.
That the spec is specific enough to hand to an agent. It reads text against a fixed vocabulary, so it cannot tell you the decision is right. Any open product choice still blocks until a person answers it.
From a repository with an `intent.md`, run `npx @pathmode/mcp-server@latest adopt` and give the printed link and code to the person connecting the repository. They choose a workspace and approve in the browser. No spec content leaves the machine until approval.
stdio, via `npx @pathmode/mcp-server`. Local mode needs no authentication; a connected workspace uses the key issued during `adopt`.

Local-first credential broker for MCP. AI coding agents (Claude Code, Cursor, Codex…) call APIs, SSH and databases through scoped presets; secrets stay in your OS keychain and never enter the agent's context.

MCP server for AI Agents — exposes Yocoolab feedback threads and design selections as tools

Know when your AI agents, cron jobs and CI pipelines fail, stall or go quiet. Your agent sets up monitors from a plain description, tells you what is broken and why, reads run history and traces, and writes notes on inci

One case file per task — shared by every session, agent and person that touches it. Self-hosted MCP server and live board. MIT.
OrangePro local-first CLI + MCP server for behavior mapping, grounded test generation, and dynamic proof.
Comments