Client Agent Workspace Templates

Status / authority: Current client-readable template companion
Effective date: 2026-08-04
Dependencies: Client_Agent_Navigation_Standard.md; Client_Agent_Workspace_Setup_v0.md; ../LEGANT_METHOD_REFERENCE.md
Supersedes: Earlier ARC-first template routing where it conflicts. Existing client artifacts remain unchanged.

Purpose: Lightweight starting structures for full-engagement, client-owned Innate Edge navigation workspaces.

Use these after reading Client_Agent_Navigation_Standard.md and Client_Agent_Workspace_Setup_v0.md.

Templates are starting points rather than required forms. Keep only what helps future agents navigate the client. Delete sections that would invite filler.

Access_Manifest.md

# Access Manifest

This file records intended access for <Client>'s folder. It is not the permission system. Google Drive sharing settings are the authority.

This is an admin map for the client, Chris, and their authorized agents. Keep it in the shared workspace when the access boundary would otherwise be unclear.

## Access Zones

| Zone | Path | Client Access | Maintainer Access | Notes |
|---|---|---|---|---|
| Shared methodology library | `Methodology Library/` | Viewer | Editor | Client-readable methodology only. |
| Client navigation workspace | `./` | Proposed: editor/commenter after Chris review | Editor | Client-owned AI context, source material, project pages, and wiki. |
| Client transcripts | `Transcripts/` | Proposed: editor | Editor | Raw client conversations only; no Chris decompressions. |
| Client evidence/source data | `Evidence/` | Proposed: editor | Editor | Psychometrics, qualitative input, and evidence extraction files. Additions should be logged in `Agent System/Source_Change_Log.md`. |
| Current deliverables | `./` | Proposed: viewer/editor depending on use | Editor | Latest official OM / Positioning files live at shared root. |
| Historical deliverables | `Historical Deliverables/` | Proposed: viewer | Editor | Prior X-Ray, Decode, old OM, old Positioning files. |
| Private maintainer context | Outside the shared client workspace | No default access unless explicitly shared | Editor | May contain private synthesis, decompressions, calibration, or strategic judgment. |
| Legant-system internals | Outside the shared client workspace | No access | Editor | Production procedures, implementation tools, and other client folders are not part of the client packet. |

## Permission Change Log

| Date | Actor | Target | Permission Change | Reason | Verified |
|---|---|---|---|---|---|
| YYYY-MM-DD |  |  |  |  |  |

AGENTS.md

# AGENTS.md -- <Client> Innate Edge

**Subject:** <Client>
**Purpose:** First file the AI reads when working inside <Client>'s client-owned Innate Edge workspace.
**Last updated:** YYYY-MM-DD

---

## Operating Objective

You are <Client>'s long-horizon alignment advisor.

Use <Client>'s Operating Manual and wiki as a provisional map, not settled truth. Help <Client> make decisions that are congruent, coherent, and contributing.

## Load Order

Before making claims or recommendations about <Client>, read:

1. `AGENTS.md`
2. `Agent System/Reference_Index.md`
3. `Wiki/Timeline.md`
4. `Agent System/Methodology_Library.md` and `Methodology Library/Navigation/Client_Agent_Navigation_Standard.md`
5. `Wiki/Current_State.md`
6. `Wiki/Personal_Operating_Manual.md`
7. `Wiki/Development_Arc.md` if present
8. `Wiki/Navigation_Log.md` when the task is meaningful navigation
9. relevant project pages in `Projects/`
10. source files only when more depth or freshness is needed

For ordinary advisory replies, follow the shared [Client Agent Reply Standard.md](https://legantlabs.com/methodology/library/protocols/client-agent-reply-standard.html), then apply client-specific preferences from `Wiki/AI_Working_Preferences.md` if present.

## Navigation Continuity

For meaningful navigation work, start or update a `Wiki/Navigation_Log.md` entry before giving substantive advice. Do not log canary tests unless Chris or the client explicitly promotes the test into real context.

Use the map -> reality -> 3C -> stepping stone -> update signal loop. Update factual context directly when source-backed, low-inference, decision-relevant, and in scope. Route self-model, strategy, positioning, and interpretation changes as open questions, update signals, or approval diffs.

Leave interrupted or unresolved episodes open rather than inventing closure.

## Access Boundary

Assume only this `<Client> Innate Edge/` folder is shared. Do not rely on parent or sibling folders unless Chris explicitly provides them.

## Source Layers

1. **Shared methodology library:** `Methodology Library/`, read-only.
2. **Client-owned source material:** current deliverables at root; historical deliverables in `Historical Deliverables/`; transcripts in `Transcripts/`; evidence/source data in `Evidence/`; notes, project files, and client-added annotations where locally mapped.
3. **Client wiki:** agent-first, client-readable AI-maintained summaries.
4. **Workspace schema:** this file, `Agent System/Reference_Index.md`, and change logs.

## Context Update Routing

Follow [Client_Agent_Navigation_Standard.md](https://legantlabs.com/methodology/library/protocols/navigation/client-agent-navigation-standard.html).

Update directly only when the claim is factual, source-backed, low-inference, decision-relevant, and in scope.

Route self-model, strategy, positioning, and interpretation changes as open questions, update signals, or approval diffs.

## Source Change Rule

When the client, Chris, or an AI changes client-owned source material, update `Agent System/Source_Change_Log.md`.

Do not overwrite current Innate Edge deliverable versions. If the Operating Manual or Positioning Report needs revision, create a dated new version or annotation file and log it.

## Source Truth Wins

If a wiki page and a source file disagree, the source file wins. Flag the mismatch and update or propose an update according to Context Update Routing.

Keep speaker attribution intact. If a frame came from Chris as a prompt, challenge, or synthesis and the client only responded to it, say that. Do not rewrite Chris-originated test language as client-originated evidence.

## Client-Specific Challenge Style

Add only the challenge patterns that actually improve guidance for this client.

## Things The AI Should Not Do

Add client-specific no-go patterns. Keep them concrete.

Reference_Index.md

# Reference Index

**Purpose:** Directory for <Client>'s client-owned AI workspace. Points LLM agents to shared methodology, client-owned sources, wiki pages, and project context.

---

## Shared Methodology Library

- `Agent System/Methodology_Library.md`
- `Methodology Library/`
- `https://legantlabs.com/methodology`: current method architecture
- `Methodology Library/LEGANT_METHOD_REFERENCE.md`: current detailed methodology route
- `Methodology Library/Navigation/Client_Agent_Navigation_Standard.md`
- `Methodology Library/Client Agent Reply Standard.md`
- `Methodology Library/ARC_Reference_Index.md`: legacy interpretation only
- `Methodology Library/ARC Methodology.docx`: legacy access notice only; complete historical source is private

Rule: use the synced `Methodology - Client Readable/` folder for client-agent work. If the client's local Google Drive sync places that folder somewhere else, use the local synced path.

## Client-Owned Sources

- `<Client>_Operating_Manual_<YYYYMMDD>.pdf`
- `<Client>_Operating_Manual_<YYYYMMDD>.docx`
- `Historical Deliverables/`
- `Transcripts/`
- `Evidence/Psychometrics/`
- `Evidence/Qualitative Input/`
- `Evidence/Evidence Extraction/`
- `Projects/`

Change tracking:

- `Agent System/Source_Change_Log.md`

## Wiki Pages

- `Wiki/Timeline.md`
- `Wiki/Current_State.md`
- `Wiki/Personal_Operating_Manual.md`
- `Wiki/Development_Arc.md`
- `Wiki/Decisions_and_Open_Questions.md`
- `Wiki/Navigation_Log.md`
- `Wiki/Conversation_Transcript_Index.md`
- `Wiki/AI_Working_Preferences.md`
- `Wiki/_changelog.md`

## Project Pages

- `Projects/<Project>/<Project>_Context_Brief.md`

## Private Maintainer Context

Private maintainer context lives outside this client workspace. Use it only when Chris explicitly shares a relevant source or asks for maintainer-facing synthesis. Never assume access from a local path or another client's workspace.

## Update Rules

- Meaningful navigation interactions start in `Wiki/Navigation_Log.md` before substantive advice.
- Factual, source-backed, low-inference, decision-relevant context can update directly with `_changelog.md`.
- Self-model, strategy, positioning, and interpretation changes route to `Decisions_and_Open_Questions.md` or an approval diff.
- Client-owned source edits require an `Agent System/Source_Change_Log.md` entry.
- Revise current deliverables through a dated new version or annotation file; never overwrite them.

Wiki/Personal_Operating_Manual.md

# Personal Operating Manual

---
**Page:** Personal Operating Manual
**Last updated:** YYYY-MM-DD
**Updated by:** <Agent> for Chris
**Purpose:** Slow map of <Client>'s durable wiring for agent-assisted navigation.
**Primary reader:** LLM agents working with <Client>'s materials; client-readable.

**Synthesized from:**
- `<source>`

**Raw-source pointers:**
- `<source>`

**Update triggers:** new wiring evidence, source contradiction, client correction, Chris correction, formal OM revision, major decision evidence.

**Approval model:** factual context may update directly when source-backed; self-model changes route as update signals or approval diffs.

**Line cap:** 150 content lines.
---

## Agent Reading Principle

Before using this page, read [Client_Agent_Navigation_Standard.md](https://legantlabs.com/methodology/library/protocols/navigation/client-agent-navigation-standard.html).

Use this question to guide the page:

> What would a future agent need to know about this client's durable wiring to help them make a more congruent, coherent, contributing decision?

This page is the durable map. Use other wiki pages for current terrain.

Because the agent does not have the client's memory, major claims should connect to examples, source pointers, or evidence summaries elsewhere in the folder. Mark unanchored claims as hypotheses, gaps, or update signals.

## Stable Wiring Summary

Name the current OM hypothesis. Keep it compact and provisional.

## Client-Facing Decision Ritual

Preserve the natural questions or moves from the client-facing OM that make the map usable in ordinary agent conversations.

## How The Verb Shows Up

Show the recurring mechanism sequence across domains. Include only examples that help future guidance.

## Engine / Fuel / Rest Boundary

Name how to tell core engine work from curiosity, play, practice, favors, maintenance, or rest.

## Operating Context For Agents

Name how to work with this client's wiring in live advisory use.

## Sources Of Scatter / Nozzle Conditions

Name the recurring drift pattern and what concentrates it.

## Boundary Tradeoffs

Name what the client is consciously giving up by honoring the boundary, not just what they should avoid.

## Language That Works / Misfires

Capture language that reliably helps or distorts navigation.

## Current Vehicle Cheatsheet

Use only for active or decision-relevant vehicles.

## Decision Filters

Include filters that help the client choose or shape commitments.

## Restlessness / Reshape Protocol

Name what an agent should check before recommending that the client stay, leave, redesign, reshape, or hand off something.

## Falsifiers / Watch Points

Name evidence that would strengthen, weaken, or refine the map.

## What Agents Should Not Do

Concrete no-go patterns only.

Wiki/Navigation_Log.md

# Navigation Log

---
**Page:** Navigation Log
**Last updated:** YYYY-MM-DD
**Updated by:** <Agent>
**Purpose:** Preserves meaningful navigation episodes so future agents can stay current with the client's decisions, doubts, recommendations, and evidence to watch.
**Primary reader:** LLM agents working with the client's materials; client-readable.

**Synthesized from:**
- [Client_Agent_Navigation_Standard.md](https://legantlabs.com/methodology/library/protocols/navigation/client-agent-navigation-standard.html)
- [Client_Agent_Workspace_Setup_v0.md](https://legantlabs.com/methodology/library/protocols/navigation/client-agent-workspace-setup-v0.html)

**Raw-source pointers:**
- `../Projects/`
- `../Transcripts/`
- `../Evidence/`
- `../Historical Deliverables/`
- `Timeline.md`
- `Current_State.md`
- `Personal_Operating_Manual.md`
- `Decisions_and_Open_Questions.md`

**Update triggers:** meaningful navigation interaction; decision, doubt, project pressure-test, self-model question, live evidence update, or interrupted navigation thread.

**Approval model:** factual episode logging can be updated directly by the client's agent; self-model, strategy, positioning, and interpretation changes route to `Decisions_and_Open_Questions.md` or an approval diff.

**Line cap:** 200 content lines; archive older closed entries when needed.
---

## Logging Rule

Start an entry before giving substantive advice on a real meaningful decision, doubt, project pressure-test, self-model question, or live evidence update.

Log episodes, not transcripts. Capture only what future agents need to avoid restarting from zero.

Promote durable facts and signals into the right wiki pages. This log is the trail, not the context layer.

Do not log canary tests unless the maintainer or client explicitly promotes the test into real context. Keep canary notes in the current chat or a maintainer-private setup/test file.

## Status Values

- `Active`: interaction is underway.
- `Closed`: recommendation, response, and context routing are captured.
- `Deferred`: decision remains open.
- `Interrupted`: session ended before closure.

## Template

```text
### YYYY-MM-DD -- [Topic]

Status:
Actor:
Platform / chat locator:
Question / doubt:
Sources read:
Agent recommendation:
Client response:
Chosen stepping stone:
Evidence to watch:
Context updates made or proposed:
Links / references:
```

## Active / Recent Episodes

No live navigation episodes logged yet.