{
  "markdown": "# ContextLattice\n\n<p align=\"center\">\n  <a href=\"https://contextlattice.io/\" target=\"_blank\" rel=\"noopener noreferrer\">\n    <img src=\"docs/readme/contextlattice-editorial-hero.png\" alt=\"ContextLattice editorial website hero showing a live local context field\" width=\"100%\" />\n  </a>\n</p>\n\n<p align=\"center\">\n  <strong>Open an agent. Already there.</strong>\n</p>\n\n<p align=\"center\">\n  The local-first intelligence layer that gives AI agents durable continuity, explainable retrieval, portable context, and verified learning across harnesses.\n</p>\n\n<p align=\"center\">\n  <a href=\"#quickstart\"><img src=\"https://img.shields.io/badge/Interface-CLI%20First-111111?style=for-the-badge\" alt=\"CLI first\"></a>\n  <a href=\"https://github.com/sheawinkler/ContextLattice/releases/tag/v5.0.2\"><img src=\"https://img.shields.io/badge/Release-v5.0.2-292929?style=for-the-badge\" alt=\"ContextLattice v5.0.2\"></a>\n  <a href=\"#quickstart\"><img src=\"https://img.shields.io/badge/Runtime-Local%20First-404040?style=for-the-badge\" alt=\"Local-first runtime\"></a>\n  <a href=\"LICENSE\"><img src=\"https://img.shields.io/badge/License-Apache%202.0-575757?style=for-the-badge\" alt=\"Apache License 2.0\"></a>\n</p>\n\n<p align=\"center\">\n  <a href=\"#quickstart\">Quickstart</a> ·\n  <a href=\"#how-it-works\">How it works</a> ·\n  <a href=\"#connect-your-agents\">Agent setup</a> ·\n  <a href=\"#architecture\">Architecture</a> ·\n  <a href=\"https://contextlattice.io/docs/\">Docs</a> ·\n  <a href=\"https://contextlattice.io/updates.html\">Updates</a>\n</p>\n\n## Stop replaying the brief\n\nModels can reason. Harnesses can act. Neither reliably retains the mission when a chat, model, tool, account, or computer changes.\n\nContextLattice gives that work a durable, inspectable context layer. It reconstructs the active objective, selects the evidence that matters, carries it safely, and records what actually worked—without turning every prompt into a transcript dump or making cloud storage mandatory.\n\n| Capability | What changes |\n| --- | --- |\n| **Durable continuity** | Reopen the objective, decisions, repository state, risks, proof, and next move as one bounded packet. |\n| **Explainable retrieval** | Rank evidence by impact per token and expose source coverage, omissions, opposition, degradation, and receipts. |\n| **Portable context** | Move signed, least-privilege continuation across agents and machines while keeping execution and transport caller-owned. |\n| **Verified skill evolution** | Discover skills without loading every file, evaluate repeated wins on holdouts, and require review before promotion. |\n| **Privacy-bounded Aggregate Signal** | Learn from explicitly opted-in, clipped statistics while raw memory remains local; production activation stays hard-blocked pending independent privacy and utility review. |\n\nThe **CLI is the primary interface**. The dashboard makes behavior and proof visible. HTTP and MCP are companion integration surfaces for applications and harnesses.\n\n## How it works\n\n| Stage | ContextLattice does |\n| --- | --- |\n| **01 · Reopen** | Reconstructs the one active mission from durable checkpoints and current state. |\n| **02 · Select** | Retrieves high-signal evidence into a compact Context Pack with provenance. |\n| **03 · Move** | Carries signed, bounded context through Agent Packets, Passports, and encrypted continuation envelopes. |\n| **04 · Earn** | Records outcomes and promotes reusable behavior only after deterministic proof and human approval. |\n| **05 · Compound** | Improves future retrieval while preserving corrections, contradictions, freshness, and retirement semantics. |\n\nContextLattice does not replace your agent harness, choose goals from retrieved text, or execute imported context. Local tools remain execution surfaces; memory and remote content remain evidence.\n\n## Quickstart\n\nRequirements: macOS, Linux, or Windows through WSL2; a Compose v2-compatible container runtime; and `gmake`, `jq`, `rg`, `python3`, and `curl`. The tested macOS baseline uses OrbStack through its explicit Docker context; see the [container runtime decision](docs/runtime/container-runtime-decision.md).\n\n### 1. Install\n\n```zsh\ngit clone https://github.com/sheawinkler/ContextLattice.git\ncd ContextLattice\ncp .env.example .env\ngmake quickstart\n```\n\n`gmake quickstart` is the prescribed technical install path; installers are bootstrap alternatives. The command prepares environment wiring, asks for a runtime profile, launches the selected local stack, and validates initial readiness.\n\n### 2. Verify the runtime and retrieval path\n\n```zsh\ncurl -fsS http://127.0.0.1:8075/health | jq\ncontextlattice doctor --pretty\ncontextlattice state status --pretty\ncontextlattice context \"verify this ContextLattice installation\" \\\n  --project contextlattice \\\n  --pretty\n```\n\nHealthy containers are only the first check. The state command verifies the canonical gateway-owned storage inventory; the final command exercises the actual context path and reports source coverage, degradation, evidence, and next actions. Existing installs can use the explicit, reversible procedure in [gateway state migration](docs/runtime/gateway-state-root.md).\n\nFor a fuller lifecycle proof from the repository:\n\n```zsh\nscripts/agent/agent-runtime-proof-pack --pretty\nscripts/agent/agent-adoption-proof-matrix \\\n  --skip-provider-smoke \\\n  --progress \\\n  --pretty\n```\n\n## Connect your agents\n\nRun integration from the repository that should use ContextLattice:\n\n```zsh\ncd /path/to/your/project\n\ncontextlattice_adopt integrate \\\n  --repo . \\\n  --agents codex,claude-code,opencode,hermes-agent,hermes-ultra,omp,mercury-agent,pi,droid \\\n  --pretty\n\ncontextlattice_adopt integrate \\\n  --repo . \\\n  --agents codex,claude-code,opencode,hermes-agent,hermes-ultra,omp,mercury-agent,pi,droid \\\n  --check \\\n  --pretty\n```\n\nThe integration command writes bounded managed blocks while preserving existing instruction text. It does **not** install Codex, Claude Code, OpenCode, Hermes, OMP, Mercury, Pi, Droid, or another third-party agent harness.\n\nExternal provider discovery is network-free. Provider execution remains explicit and caller-authorized; see the [external-provider boundary](docs/runtime/external-provider-auth-boundary.md).\n\nIf an agent is performing the installation, it should follow the quickstart directly, avoid cloning a second checkout when already inside one, report the exact failing command and path, and rerun the deterministic check after any repair.\n\n## The daily agent loop\n\n```zsh\n# Confirm readiness when the environment is uncertain.\ncontextlattice doctor --pretty\n\n# Retrieve scoped context before substantial work.\ncontextlattice context \"debug the current release regression\" \\\n  --project contextlattice \\\n  --pretty\n\n# Save concise, durable progress.\ncontextlattice remember \\\n  \"Root cause verified; regression test added; focused checks pass.\" \\\n  --project contextlattice \\\n  --pretty\n\n# Resume without replaying the transcript.\ncontextlattice resume --project contextlattice --pretty\n\n# Repair stale or wrong recall without silently rewriting history.\ncontextlattice correct \\\n  \"The prior deployment record is stale.\" \\\n  --category stale \\\n  --project contextlattice \\\n  --pretty\n\n# Close the loop with the verified outcome.\ncontextlattice finish \\\n  \"Regression fixed and verified.\" \\\n  --success \\\n  --project contextlattice \\\n  --pretty\n\n# Project the next bounded move or bind a completed response to durable proof.\ncontextlattice_continuous_cognition status \"prepare the next verified move\" \\\n  --project contextlattice --session-id <session-id> --agent-id codex_gpt5 \\\n  --task-id <task-id> --objective-id <objective-id> --as-of <rfc3339> --pretty\ncontextlattice_continuous_cognition evaluate \"verify the completed response\" \\\n  --project contextlattice --session-id <session-id> --agent-id codex_gpt5 \\\n  --task-id <task-id> --task-identity-id <task-identity-id> --as-of <rfc3339> --pretty\n\n# Prepare context for an external worker without exposing its one-shot claim.\ncontextlattice agent-fit context-prep-schedule --project contextlattice \\\n  --session-id <session-id> --agent-id codex_gpt5 --payload-file prep-request.json --raw\ncontextlattice agent-fit context-prep-claim --project contextlattice \\\n  --session-id <session-id> --agent-id codex_gpt5 --prep-id <prep-id> \\\n  --worker-id <worker-id> --claim-token-file prep.claim --raw\ncontextlattice agent-fit context-prep-complete --project contextlattice \\\n  --session-id <session-id> --agent-id codex_gpt5 --prep-id <prep-id> \\\n  --claim-token-file prep.claim --payload-file prep-artifact.json --raw\ncontextlattice agent-fit context-prep-use --project contextlattice \\\n  --session-id <session-id> --agent-id codex_gpt5 --prep-id <prep-id> \\\n  --task-id <task-id> --effective-profile-digest <sha256-digest> \\\n  --source-generation <generation> --raw\n```\n\nContinuous Cognition is advisory-only: each invocation makes one bounded\nrequest, returns opaque evidence references, and never dispatches a runner or\nperforms an external mutation. Context-preparation claims stay in an owner-only\nfile and cross the completion/failure boundary only through the protected\nheader; successful explicit use consumes the artifact once.\n\nFind a capability without loading every skill body:\n\n```zsh\ncontextlattice_skills_index search \"browser automation\" --pretty\n```\n\nThe active Skills Index scans configured Codex, Hermes, Hermes Ultra, and shared\nagent roots read-only. It reports each harness and root inventory separately,\ncollapses byte-identical `SKILL.md` files by SHA-256 digest while retaining every\nsource path as provenance, and requires discriminating query-term coverage\ninstead of ranking generic words such as `skill`, `index`, or `agent`.\nQuarantine discovery remains separate, read-only by default, and never\nauto-promotes retrieved content.\n\n## Architecture\n\n<table>\n  <tr>\n    <td width=\"50%\">\n      <a href=\"https://contextlattice.io/architecture.html\">\n        <img src=\"docs/public_overview/assets/architecture-service-map.svg\" alt=\"ContextLattice service map\" width=\"100%\" />\n      </a>\n    </td>\n    <td width=\"50%\">\n      <a href=\"https://contextlattice.io/architecture.html\">\n        <img src=\"docs/public_overview/assets/architecture-retrieval-flow.svg\" alt=\"ContextLattice retrieval and learning flow\" width=\"100%\" />\n      </a>\n    </td>\n  </tr>\n</table>\n\nThe default local control path is:\n\n```text\nAgent or application\n        │\n        ▼\nContextLattice CLI / HTTP / MCP\n        │\n        ▼\nGateway :8075\n        ├── durable write and outbox fanout\n        ├── scoped retrieval and source receipts\n        ├── session, objective, graph, and outcome state\n        └── dashboard-visible proof and operations\n```\n\nWrites are validated and durably persisted before fanout. Retrieval merges the available sources, ranks bounded evidence, and reports missing or degraded coverage instead of hiding it.\n\nThe active application path is Go and Rust. Python remains in build, development, migration, and audit tooling rather than the live request path. The exact runtime and toolset choices are recorded in the [v4 runtime decision](docs/runtime/v4-runtime-toolset-decision.md) and [container decision](docs/runtime/container-runtime-decision.md).\n\n## Public and paid boundaries\n\nThe public local lane is account-free and useful on its own. It includes the CLI-first memory lifecycle, Context Packs, sessions, graph and claim surfaces, Skills Index discovery, Agent Packets, public Passport and Mesh contracts, and local proof tooling.\n\nPaid artifacts add governed collaboration, protected activation, workspace operations, advanced analytics, and hosted distribution. They do not turn local memory into a mandatory cloud dependency.\n\nSee [plans and distribution boundaries](docs/public_overview/premium.html) for the current contract.\n\n## Install options\n\nmacOS technical preview: unsigned DMG bootstrap launcher; expect Gatekeeper warnings until Developer ID notarization is configured, and prefer the source/CLI path.\n\n| Path | Best for | Status |\n| --- | --- | --- |\n| Source + `gmake quickstart` | Technical users and terminal-capable agents | **Recommended** |\n| `brew tap sheawinkler/contextlattice && brew install --cask contextlattice` | macOS convenience bootstrap | Available |\n| [macOS universal DMG](https://github.com/sheawinkler/ContextLattice/releases/latest/download/ContextLattice-macOS-universal.dmg) | Guided macOS bootstrap | Unsigned technical preview; expect Gatekeeper warnings |\n| [Windows x64 MSI](https://github.com/sheawinkler/ContextLattice/releases/latest/download/ContextLattice-windows-x64.msi) | Guided Windows bootstrap | Available |\n| [Linux bootstrap bundle](https://github.com/sheawinkler/ContextLattice/releases/latest/download/ContextLattice-linux-bootstrap.tar.gz) | Guided Linux bootstrap | Available |\n\n### Resource profiles\n\n| Profile | CPU | RAM | Storage |\n| --- | --- | --- | --- |\n| Hugging Face / Glama lite | `2–4` vCPU | `4–8 GB` | `20–50 GB` SSD |\n| Local Lite core | `2–4` vCPU | `8–12 GB` | `25–80 GB` SSD |\n| Local Lite advanced | `4–6` vCPU | `12–16 GB` | `80–140 GB` SSD |\n| Local Full | `6–8` vCPU | `12–20 GB` | `100–180 GB` SSD |\n\nFor heavier ingest, model storage, or the spike-lab adapters, read the [installation and storage guidance](https://contextlattice.io/installation.html) before selecting a profile.\n\n## Security and privacy\n\n- Local-first and account-free in the public local lane.\n- API-key protection for operational routes.\n- Deterministic secret-like content filtering at write ingress: redact by default, block when configured, and allow only by explicit operator choice.\n- Provenance and trust isolation on retrieved memory.\n- Signed portable context and encrypted continuation envelopes.\n- Dry-run-first graph repair, source backfill, and quarantine workflows.\n- No automatic execution of retrieved instructions or imported continuation content.\n\nSecurity reports follow [SECURITY.md](SECURITY.md).\n\n## Documentation\n\n| Need | Start here |\n| --- | --- |\n| Product overview | [contextlattice.io](https://contextlattice.io/) |\n| Installation | [Installation guide](https://contextlattice.io/installation.html) |\n| CLI and agent lifecycle | [CLI reference](https://contextlattice.io/cli.html) |\n| Harness and app integration | [Integration guide](https://contextlattice.io/integration.html) |\n| Architecture and scaling | [Architecture](https://contextlattice.io/architecture.html) · [Scaling memory](https://contextlattice.io/scaling-memory.html) |\n| Troubleshooting | [Troubleshooting guide](https://contextlattice.io/troubleshooting.html) |\n| Current behavior and release evidence | [Updates](https://contextlattice.io/updates.html) · [v5.0.2 release notes](docs/releases/v5.0.2.md) |\n| Roadmap | [Public roadmap](https://contextlattice.io/roadmap.html) |\n| Agent hooks | [Agent hook contract](docs/agent-hooks.md) |\n| Retrieval trust | [Retrieval receipts](docs/retrieval-receipts.md) |\n| Skills and verified learning | [Skill efficacy review](docs/skill-efficacy-review.md) · [Skill Foundry](docs/outcome-policy-skill-foundry.md) |\n| Portable context | [Context Passport and Mesh](docs/context-passport-mesh.md) |\n| Local inference | [Local model options](docs/runtime/local-model-options.md) |\n| Full repository-backed manual | [Public field manual](docs/wiki/README.md) |\n\nThe current release baseline is [`v5.0.2`](https://github.com/sheawinkler/ContextLattice/releases/tag/v5.0.2).\n\n## License\n\nContextLattice's public lane is licensed under the [Apache License 2.0](LICENSE).\n",
  "bytes": 15531,
  "sha": "5065a28557e2d32e472f37ca10f13dccc1afd93b62acdaf4f389671aad2a1d86",
  "repo_slug": "sheawinkler/contextlattice",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_sheawinkler_contextlattice_287e1d97/readme"
}