TLDRGraph

TLDRGraph turns a repository into a source-backed feature catalog and interactive
Workflow Explorer. Your coding agent reads the repository, identifies meaningful
product and technical capabilities, and records every displayed step with file,
symbol, and line-range evidence.

TLDRGraph does not run an AI process itself and does not infer workflows from a
static graph. It coordinates direct catalog and workflow artifacts with the
coding agent that already has your repository open.

Install

pip install tldrgraph

Generate a catalog

Run this inside a supported coding-agent session:

tldrgraph init

When init reports needs_feature_workflows, it includes the current source
hash. The active agent identifies feature outcomes and immediately writes the v4
.tldrgraph/features.yaml index. It delegates one indexed feature to each
source-reading subagent; each worker writes only its own v4 workflow file. Run
tldrgraph init again to validate the artifacts and generate the explorer.

When repository source changes after a catalog exists, run tldrgraph refresh
instead. It performs the same validation and generation flow while making the
update intent explicit.

The final artifacts are:

• .tldrgraph/features.yaml
• .tldrgraph/workflows/<feature_id>.yaml
• .tldrgraph/TLDRGRAPH_VISUALIZER.html

Explore

tldrgraph ui --serve

The standalone browser application groups product and technical capabilities,
draws each proven workflow, and opens cited source ranges from the repository.

CLI

tldrgraph init [PATH] [--json]
tldrgraph refresh [PATH] [--json]
tldrgraph ui [--path PATH] [--serve] [--port PORT] [--open|--no-open]
tldrgraph install [--path PATH] [--all-agents]

Version 0.3 is a breaking, workflow-only release. Earlier graph-based commands
and v1/v2 generated artifacts are not supported; run tldrgraph init to create
the current catalog, then tldrgraph refresh after later source changes.