agentleFS
PhoenixNestPhoenixNest/AGENT-GLOBAL-BASE/telescope/CLAUDE.md

AGENT-GLOBAL-BASE / telescope

CLAUDE.md2 starsChanged 2 months agoApache-2.0View on GitHub

Cross-department pointer for the workspace's three department research archives, plus the home for research that is genuinely workspace-wide in scope. Read this before documenting any investigation, or before deciding which department's telescope a new investigation belongs in. Research direction is documented separately per department, each with its own dedicated archive: This folder is not a general document repository. It holds only: Applied at the time a research investigation is commissioned: 1. Default: file it under the requesting or most-affected department's…

The file

We split this CLAUDE.md into the rules it gives an agent

8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Cross-cutting exception: if the investigation's findings affect workspace-wide governance or shared configuration used by every department …

    line 33
  2. Distinctive:

    When a cross-cutting report is filed here, cross-link it from any department instance's index it materially concerns, so it isn't siloed …

    line 37
  3. Very specific:

    Only create a report directly in this folder if it meets the cross-cutting exception above.

    line 59
  4. Very specific:

    Create a new folder: YYYY-MM-DD-<slug>/

    line 62
  5. Distinctive:

    Add an entry to README.md (the cross-department index) and cross-link from any department index it concerns

    line 65
  6. Very specific:

    Every report lives in a YYYY-MM-DD-<slug>/ folder containing research-report.md.

    line 103
  7. Distinctive:

    Every report is indexed in its instance's README.md.

    line 104
  8. Distinctive:

    Use the shared template — do not invent alternative formats.

    line 105

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

telescope/ — Cross-Department Research Index

Cross-department pointer for the workspace's three department research archives, plus the home for research that is genuinely workspace-wide in scope. Read this before documenting any investigation, or before deciding which department's telescope a new investigation belongs in.


What This Is

Research direction is documented separately per department, each with its own dedicated archive:

InstanceScopeOwner
core-component-00/telescope/Engineering + cutting-edge LLM researchDr. Elias Vance, Lab Director
company/telescope/Product-oriented researchMarcus Tran-Yoshida, CPO
studio/casual-games/telescope/Game-design / competitive / live-ops / market researchMarcus Vogel, Studio Director
telescope/ (this folder)Cross-department index + workspace-wide governance researchOrganizer / CEO

This folder is not a general document repository. It holds only:

  1. This index/pointer documentation
  2. Research that is genuinely cross-cutting — see the classification rule below
  3. The canonical template/ (each department instance also keeps its own copy)

Classification Rule — Where Does a New Investigation Go?

Applied at the time a research investigation is commissioned:

  1. Default: file it under the requesting or most-affected department's instance.
  2. Cross-cutting exception: if the investigation's findings affect workspace-wide governance or shared configuration used by every department (e.g. .claude/ config, a workspace-wide rule like mcp-governance.md, or cross-cutting infrastructure no single department owns), it stays here at the workspace root instead.
  3. When a cross-cutting report is filed here, cross-link it from any department instance's index it materially concerns, so it isn't siloed from departments that rely on its findings.

Directory Structure

telescope/
├── README.md              ← Cross-department index — see this for the full picture
└── template/               ← Canonical report template (each department instance mirrors this)
    ├── research-report.md
    └── qa-document.md

A cross-cutting investigation gets its own YYYY-MM-DD-<slug>/ folder under this root per the rule above.


Creating a New Research Report Here

Only create a report directly in this folder if it meets the cross-cutting exception above. Otherwise, go to the relevant department's telescope/ instance and follow its CLAUDE.md.

  1. Create a new folder: YYYY-MM-DD-<slug>/
  2. Copy the template: template/research-report.md → YYYY-MM-DD-<slug>/research-report.md
  3. Complete the report in the new folder
  4. Add an entry to README.md (the cross-department index) and cross-link from any department index it concerns

Telescope Conventions

These rules apply to every telescope instance — root and all three departments.

Report Shape — two canonical forms

  • Simple — a single research-report.md (+ optional qa-document.md) at the investigation folder root. Nothing else. Use this for any self-contained investigation.
  • Programme — a top-level research-report.md (required; the consolidated synthesis) and optional qa-document.md, plus exactly one supporting/ folder holding everything ancillary: diagrams, deployment/migration plans, sub-investigation reports, implementation plans. No other subfolder name is permitted at the investigation root. A sub-investigation's own report still lives at supporting/<sub-slug>/research-report.md, and stays fully template-conformant. No inventing new folder names per report.

Status Lifecycle — four states

In Progress / Complete / Superseded / Abandoned. An investigation's Status must be updated in both its own research-report.md metadata table and its instance README.md index entry in the same commit — the README is the authoritative "what's active" view; do not rely on opening every folder to find out. Any In Progress investigation untouched for 90+ days must be reviewed by the instance owner (Lab Director / CPO / Studio Director-or-delegate) and reclassified as Complete, Abandoned, or recommitted with a new note — it may not sit silently as "active" indefinitely.

No physical archive//running/ subfolders. Status-based physical reorganization forces a file move on every status change, which fights the append-only rule below and churns git history without improving retrieval — the README Status column already does this job.


Rules (unchanged across all telescope instances)

  • Every report lives in a YYYY-MM-DD-<slug>/ folder containing research-report.md.
  • Every report is indexed in its instance's README.md.
  • Use the shared template — do not invent alternative formats.
  • This archive is append-oriented. Do not edit published reports except to correct factual errors (note the correction with a date in the report's metadata), and do not delete completed research reports without explicit direction from the User.
  • Template skeleton (Metadata, Executive Summary, Investigation Scope, Research Questions, Methodology, Findings, Analysis, Recommendations, References, optional Appendices/Research Log/Open Questions) is fixed across all instances for citation and ADR-linkage consistency. Departments may add optional domain-specific addenda without forking the core structure.

Related Documentation

DocumentPurpose
telescope/README.mdFull cross-department index
core-component-00/telescope/CLAUDE.mdLab research archive charter
company/telescope/CLAUDE.mdCompany research archive charter
studio/casual-games/telescope/CLAUDE.mdStudio research archive charter
.claude/rules/mcp-governance.mdWorkspace MCP governance policy

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.