ContextLattice Continuity Layer

What does this agent do, and when should you use it?

ContextLattice is a local-first context service that lets agents save progress, resume work, and retrieve relevant evidence through its CLI. It organizes durable state into Context Packs and exposes session, objective, graph, and outcome surfaces. The CLI is the primary interface; a Gateway provides local HTTP and MCP integration, while a dashboard makes behavior and proof visible. Installation launches a selected local stack with a Compose v2-compatible container runtime; the active application path is written in Go and Rust. Its public local lane works without an account, while paid offerings cover collaboration, protected activation, workspace operations, and hosted distribution.

An operator can run contextlattice doctor to check readiness, then use contextlattice context to retrieve project-scoped evidence. Retrieval returns a compact, sourced context packet and reports coverage, omissions, or degradation. Agents can use remember to save progress, resume to reopen work, correct to mark stale or wrong recall, and finish to record a verified outcome. The Skills Index searches with contextlattice_skills_index search without loading every skill file. Agent Packets, Passports, and encrypted continuation envelopes carry bounded context; outcomes are recorded, and reusable behavior is promoted only after deterministic proof and human approval. ContextLattice does not execute imported context or choose goals from retrieved text.

1. An engineer switching chats, models, or computers can use resume to recover a task’s objective, decisions, and next move.
2. A developer investigating a regression can retrieve project-scoped evidence, save verified progress, and record the fix.
3. A team maintaining several agent harnesses can use contextlattice_adopt integrate to add bounded managed configuration blocks.
4. An operator who needs to audit recall can inspect source coverage, omissions, opposition, degradation, and receipts.
5. A team handing work between agents or machines can carry bounded context in Agent Packets, Passports, and encrypted continuation envelopes.

How do you install or deploy this agent?

Requirements are macOS, Linux, or Windows through WSL2; a Compose v2-compatible container runtime; and gmake, jq, rg, python3, and curl. The documented quickstart clones the repository and launches a local stack:

git clone https://github.com/sheawinkler/ContextLattice.git
cd ContextLattice
cp .env.example .env
gmake quickstart

The README does not list required API credentials; setup asks the user to select a runtime profile.

How do you use this agent?

Check runtime and retrieval readiness:

curl -fsS http://127.0.0.1:8075/health | jq
contextlattice doctor --pretty
contextlattice state status --pretty
contextlattice context "verify this ContextLattice installation" \
  --project contextlattice \
  --pretty

Save and resume work:

contextlattice remember \
  "Root cause verified; regression test added; focused checks pass." \
  --project contextlattice \
  --pretty
contextlattice resume --project contextlattice --pretty
contextlattice finish \
  "Regression fixed and verified." \
  --success \
  --project contextlattice \
  --pretty

Integrate supported harnesses from the target repository:

contextlattice_adopt integrate \
  --repo . \
  --agents codex,claude-code,opencode,hermes-agent,hermes-ultra,omp,mercury-agent,pi,droid \
  --pretty

What are this agent's strengths and limitations?

How does this agent compare with similar options?

Key facts side by side with the most closely related agents.

How does FollowAgents rate this agent?

FAQ