One path. Follow it top to bottom; it is the same path README.md, START-HERE.md,
and a generated workspace's own README describe.
Prerequisites: git, bash, awk, openssl, and python3 on PATH under that
exact name, plus an AI coding agent — Claude Code, or VS Code with GitHub Copilot.
Nothing here compiles evidence on its own; the compile step is a prompt the agent runs.
No API key of its own, no telemetry.
git clone https://github.com/FrancyJGLisboa/context-compiler-bootstrap bootstrap
./bootstrap/scripts/create-context-compiler.sh ./my-context
cd ./my-context
./scripts/preflight.sh # must print "Ready." before you go further./scripts/stage-northstar.sh .Open AI-WORKSPACE.code-workspace and ask the AI:
Update my context, then prepare a brief for Northstar Feeds.
Then, to see the rest of the surface:
What changed since June 1?
Why is BRL/USD 5.70 the current assumption?
What did we believe on June 30?
Show me only what needs review.
Northstar Feeds is fictional — every person, number, and event is synthetic
(../benchmarks/northstar/README.md).
Add evidence however is easiest — choose files or folders, paste text, paste one or
many links, or drop files into EVIDENCE-INBOX/ — then say:
Add this evidence and update my context.
Then ask for the work product: Prepare me for the Northstar meeting, What changed
since ?, Why do we think X?, What did we believe on ?, Show
me only what needs my judgment. Results are saved under BRIEFS/ and REVIEWS/.
The stable workflow behind those sentences is /ctx-add; /ctx-extract and
/ctx-compile are advanced internal controls. You do not need AGENTS.md for any of
the above — it is the schema, written for the AI tool.
One ships: client-decision (activated for you by step 2). To teach the workspace a
different kind of work, choose Create Specialization in the Context Workspace
sidebar, or just say:
Create a specialization for tracking project decisions, alternatives, owners, constraints, supersession, and unresolved risks.
The agent asks at most five domain questions, builds and tests the profile, previews observable results, and requests explicit approval before activation. It never asks you to edit schemas, run tests, or use Git.
- The original evidence is preserved.
- Supported formats are normalized; unfamiliar formats are retained and marked for attention.
- Only new or changed evidence is compiled.
- Provenance, temporal state, and profile rules are validated.
- A scoped local Git checkpoint is created for compiler-owned paths only.
No automatic push occurs. A failed checkpoint never discards successfully compiled
local output. Checkpoints stay silent until git config user.name and
git config user.email are set in the generated repo.
./scripts/preflight.sh
./scripts/verify-guided-workspace.shBoth are also on the VS Code Command Palette under Tasks: Run Task.
Install Git for Windows for Git Bash, git,
bash, awk, and openssl. Install Python 3 and make sure it is on PATH as python3
— the compiler invokes that exact name. Open the generated .code-workspace in VS
Code; compiler scripts run through Git Bash. The wiki -> context compatibility
symlink may not be created on Windows; use context/ directly if wiki/ is missing.
A local sidebar for the same actions, using the GitHub Copilot subscription already in
VS Code. It stores no credentials, sends no telemetry, and does not call a separate
model API. It is not on the Marketplace — you build the VSIX yourself, which
additionally needs Node 20+, npm, network access, and the code CLI:
./scripts/package-vscode-extension.sh # in the bootstrap clone
code --install-extension dist/context-workspace-0.1.0.vsixSee VSCODE-EXTENSION.md.
For automation, packaging, visualization, and MCP access, open
../ADVANCED.md.