Soot DocsThe practical guideMarkdown ↗

FROM INTENTION TO EVIDENCE

Work with one small Soot

Start with one clear job. Give this prompt to your coding agent, then use the guide to inspect its decisions rather than trusting a vague “done.”

Agent prompt
Read https://trysoot.com/docs/quick-start/index.md and its relevant references. First inspect the user's existing Soot installation, version, workspace/config and your permissions; do not overwrite configuration. If an installer, provider, Pack or backend is missing, identify the exact prerequisite and how an operator can prepare it. Agree ONE measurable outcome; create ONE focused Soot with a clear mission, acceptance criteria and minimum scopes, starting read-only. Validate → save → activate → verify the active revision; connect Nest to the intended runtime. Run a real permitted Test using a real provider, report Data Pool source URI/ID/current revision, or file/Skill filepath plus reviewed Git commit/hash/version, tool effects, run/config revision, outcome and a denied-scope check; demo/scripted is not AI verification. Refer to credentials only by operator-managed private file/store paths, never values in chat or argv. Do not provision infrastructure, extend grants or publish. For optional deployment, present source/image, state, secret references, costs and a backup/rollback plan for review before infrastructure effects. Reconcile uncertain outcomes with the original receipt/request/run IDs, never blindly resubmit.

One path, two sessions

INITIALIZATION turns an intention into one Soot with a mission, sources and explicit permissions. OPERATION turns runs into visible, repeatable work you can review and improve. You do not need a swarm or a VM to start.

  1. Quick start: intention to a real Test — the continuous end-to-end walkthrough.
  2. Initialization — workspace choice, mission, validation and activation.
  3. Operation — task briefs, Nest and evidence review.
  4. Data Pool — shared sources, separate permissions.
  5. Automatic collaboration — one conversation, permitted specialists and reusable group templates.
  6. Optional deployment — after local verification.
  7. Reference & troubleshooting — terms, symptoms and maintained sources.

Small responsibility, clear ownership

A Soot owns one focused responsibility. The runtime host owns policy, providers/models, vaults, limits, tool bindings and durable state. Nest is the operator workspace for conversations, management and observation. Soot Forge distributes Packs composing Skills and tools; a Pack name is not permission to execute.

  1. You
  2. Nest / CLI
  3. Soot runtime
  1. Soot runtime
  2. Model call
  3. Host-selected provider
  1. Soot runtime
  2. Selected Skill + granted tool
  3. Source / Data Pool

Host owns policy, tool bindings, provider keys and durable conversations.

Responsibility illustration · not a product screenshot
Read the text diagram
You → Nest / CLI → Soot runtime
Soot runtime → Model call → Host-selected provider
Soot runtime → Selected Skill + granted tool → Source / Data Pool
Host owns policy, tool bindings, provider keys and durable conversations.

This is an illustrative responsibility diagram, not a product screenshot or proof that services are installed. Installing ≠ granting ≠ backend start. A workspace and host shell are not an OS sandbox.

What will you finish with?

One Soot reads an approved source, produces cited drafts, passes a real-model Test and is denied outside its scope. You know which active revision produced the result and which prerequisites still need an operator. That is stronger evidence than a healthy container with no provider call.

Content reviewed: 2026-10-09