pulse8-ai-cortex-knowledge-vault
AGENTS.md15 starsChanged 5 months agoApache-2.0View on GitHub
Agent-native knowledge OS on Markdown: typed graph (NetworkX), full-text search (QMD), MarkItDown compiler, MCP + REST. No database — everything is files under the vault directory. Two naming layers: scripts/env_check.sh applies defaults, resolves CORTEX_* aliases, and writes .env for Docker. See .env.example and cortex/config.py for the full list. Never commit .env or secrets. Load the master skill first, then the sub-skill for your task:
- Reads credentials
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository10 other files it gives its agents.
- cortex-apiSkill
- cortex-authSkill
- cortex-compilerSkill
- cortex-contributingSkill
- cortex-deploySkill
- cortex-feedbackSkill
- cortex-graph-searchSkill
- cortex-mcpSkill
- cortexSkill
- cortex-vaultSkill
The file
We split this AGENTS.md into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Very specific: line 88
Prefer minimal, focused diffs.
- Very specific: line 91
Do not commit .env, API keys, or webhook URLs.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AGENTS.md — PULSE8.ai Cortex
Agent-native knowledge OS on Markdown: typed graph (NetworkX), full-text search (QMD), MarkItDown compiler, MCP + REST. No database — everything is files under the vault directory.
Repository map
| Path | Purpose |
cortex/ | Python application (FastAPI, MCP, graph, search, compiler, auth) |
scripts/ | start.sh, stop.sh, serve.py, bulk_ingest.sh, env_check.sh |
docker/ | QMD container (docker/qmd/) |
tests/ | Pytest suite |
example_vault/ | Sample vault for local dev |
docs/ | Deployment guides, design specs |
.cursor/skills/ | Project Cursor skills (load for deep workflows) |
Environment variables
Two naming layers:
| Layer | Example | Used by |
.env / Compose | VAULT_DIR, VAULT_RAW_DIR, VAULT_WIKI_DIR, LLM_API_KEY, AUTH_METHOD | scripts/start.sh, docker-compose.yml |
Python CORTEX_* | CORTEX_VAULT_PATH, CORTEX_VAULT_RAW_DIR, CORTEX_VAULT_WIKI_DIR, CORTEX_LLM_API_KEY | cortex/config.py (CortexSettings) |
scripts/env_check.sh applies defaults, resolves CORTEX_* aliases, and writes .env for Docker.
See .env.example and cortex/config.py for the full list. Never commit .env or secrets.
Essential commands
# Dependencies
uv sync --all-extras
# Start stack (QMD + Cortex)
./scripts/start.sh
./scripts/stop.sh
# Cortex only (external QMD)
./scripts/start.sh --cortex-only
# Dev server without Docker
CORTEX_MCP_TRANSPORT=http CORTEX_VAULT_PATH=./example_vault uv run python scripts/serve.py
# Tests
uv run pytest tests/ -v
# Bulk ingest (no server required)
./scripts/bulk_ingest.sh ./my-papers/Network endpoints
| Service | URL |
| Cortex REST | http://localhost:8420/api/v1/ |
| Cortex MCP | http://localhost:8420/mcp/ |
| QMD search | http://localhost:3100/ |
| OpenAPI | http://localhost:8420/docs |
Cursor skills
Load the master skill first, then the sub-skill for your task:
| Skill | When to load |
cortex | Any work in this repo; routes to sub-skills |
cortex-deploy | Docker, EC2, start.sh, logs, production ops |
cortex-vault | Vault folders, notes, frontmatter, .cortex/ |
cortex-mcp | MCP tools, Cursor/Claude Desktop, streamable HTTP |
cortex-api | REST endpoints, curl, FastAPI routes |
cortex-compiler | Ingest, compile, bulk ingest, manifests |
cortex-graph-search | Graph edges, QMD search, vault_context |
cortex-auth | API key, Entra ID OIDC, JWT |
cortex-feedback | Feedback notes, Teams notifications |
cortex-contributing | Python changes, tests, new MCP tools |
Documentation
- README.md — user-facing setup and features
- .env.example — configuration template
- docs/ec2-gpu-setup.md — GPU production on EC2
- docs/superpowers/specs/2026-05-29-cursor-skills-design.md — skills pack design
Conventions for agents
- Prefer minimal, focused diffs; match existing patterns in
cortex/. - REST and MCP share handlers in
cortex/mcp/tools.py— keep behavior in sync. - Vault paths are relative to vault root (e.g.
wiki/foo.md, not absolute paths). - Do not commit
.env, API keys, or webhook URLs.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.