Local MCP memory layer for AI coding agents
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"zihangzzz-polaris-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# Polaris Polaris stops your AI coding agent from repeating the same mistakes. Open source. Pay $2.49 once if it saves you a loop. ## Install git clone https://github.com/zhaozihangbehappy-gif/Polaris.git && cd Polaris && python3 -m pip install -r adapters/mcp-polaris/requirements.txt Optional CLI install from PyPI: `pipx install zihangzzz-polaris-mcp` ## Pay https://<YOUR_HANDLE>.gumroad.com/l/polaris One payment. Forever. The receipt email is your proof of support. No account, no login. ## Contribute Polaris grows through the people who use it. If Polaris saved you a loop — or if you hit a failure it didn't know about — you can add it: - `python3 scripts/polaris_community.py submit <your-pattern.json>` — propose a new candidate. Accepts either a full v4 pattern record or a `schema_version: 1` contribution file exported by `polaris_cli.py experience contribute` (auto-converted). - `python3 scripts/polaris_community.py confirm <pattern_id>` — tell the project a candidate actually helped you on a real case. - `python3 scripts/polaris_community.py reject <pattern_id>` — tell the project a candidate was wrong or harmful. A candidate joins Polaris's shared library after **≥ 2 independent users** confirm it helped them and no one rejected it. The shared library has three tiers — `official` (internal verified-live), `community` (community-verified through this channel), and `candidate` (unconfirmed but shipped so users can try them). All three are loaded at lookup time; each match carries its tier. Nobody's agent runs through a central server; confirmations are opt-in. See `community/README.md` for how trust and promotion work. The shipped `candidate` tier now covers all 8 ecosystems. The `official` tier remains the audited baseline, while candidate breadth grows faster through releases and community submissions. ## Read these, not the rest - START_HERE.md - FACTS.md - INSTALL.md There's an `archive/` folder too. It's old audit work I kept public so anyon…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.