Skip to content

docs: Restructure README around agent-output-quality framing#174

Open
jlevy wants to merge 6 commits into
mainfrom
claude/readme-restructure-docs-nlzk66
Open

docs: Restructure README around agent-output-quality framing#174
jlevy wants to merge 6 commits into
mainfrom
claude/readme-restructure-docs-nlzk66

Conversation

@jlevy

@jlevy jlevy commented Jun 13, 2026

Copy link
Copy Markdown
Owner

Summary

Rewrites the README to reflect how tbd works now: not just an issue tracker, but a managed-document system whose purpose is improving the quality of agent output via three levers, better knowledge, better processes, and better task management.

The previous README led with beads as "the" feature and treated specs, guidelines, and shortcuts as add-ons. This restructures around a clearer arc.

New arc

  • Quality spine: open with the single goal (better agent output) and its three levers.
  • Meta-skill surface, stated early: you talk to your agent; it routes through tbd's own docs.
  • Three pillars: Knowledge, Shortcuts (reusable processes), Tasks, delivered by four artifacts (guidelines, shortcuts, templates, beads).
  • Design principles ("Why It Works"): the pieces compose; everything is available at scale via attention-routing across dozens of focused skills; guidance is high-level rather than over-prescriptive (so it helps strong models like Opus 4.8 and Fable instead of constraining them); batteries-included yet fully forkable; git-native underneath.

Cleanup

  • Promote managed/forkable docs (tbd docs fork/update/add) from a footnote to a named section.
  • Consolidate points that were repeated (yours-to-change, scope/non-goals, "agent-written") and trim historical asides, the coordination-tools list, and the frozen tbd list dump.
  • Fix stale counts: 25+ guidelines, 30+ shortcuts, 4 templates; add the missing guideline rows and lead with the load-first set.
  • Apply strict common-doc-guidelines punctuation: no em dashes.

Diff is README-only. Work tracked under bead epic tbd-5c33.

https://claude.ai/code/session_01RDZ8fshcw3u6tqQywZBTPB

claude added 3 commits June 13, 2026 16:47
Reframe the README to reflect how tbd works now: a managed-document
system whose purpose is improving agent output quality via three levers
— better knowledge, better processes (shortcuts + templates), and better
task management — not just an issue tracker.

- New arc: quality spine -> meta-skill surface -> three pillars
  (Knowledge/Shortcuts/Tasks, four artifacts incl templates) -> design
  principles (composable + attention-routing at scale; high-level not
  over-prescriptive; batteries-included-yet-forkable; git-native; scope)
- Promote managed/forkable docs (tbd docs fork/update/add) to a named
  section instead of a footnote
- Consolidate points said multiple times (yours-to-change, scope/
  non-goals, agent-written) and trim historical asides, the
  coordination-tools list, and the frozen tbd-list dump
- Fix stale counts: 25+ guidelines, 30+ shortcuts, 4 templates; add the
  missing guideline rows and lead with the load-first set

https://claude.ai/code/session_01RDZ8fshcw3u6tqQywZBTPB
Replace all em dashes with commas, colons, semicolons, or parentheses
as appropriate, enforcing the strict no-em-dash punctuation adopted
upstream. Keeps the restructured content; only punctuation changes.

https://claude.ai/code/session_01RDZ8fshcw3u6tqQywZBTPB
@deepsource-io

deepsource-io Bot commented Jun 13, 2026

Copy link
Copy Markdown

DeepSource Code Review

We reviewed changes in b68dc4e...531806f on this pull request. Below is the summary for the review, and you can see the individual issues we found as inline review comments.

See full review on DeepSource ↗

PR Report Card

Overall Grade   Security  

Reliability  

Complexity  

Hygiene  

Code Review Summary

Analyzer Status Updated (UTC) Details
Secrets Jun 13, 2026 7:08p.m. Review ↗

Important

AI Review is run only on demand for your team. We're only showing results of static analysis review right now. To trigger AI Review, comment @deepsourcebot review on this thread.

@github-actions

github-actions Bot commented Jun 13, 2026

Copy link
Copy Markdown

Coverage Report for packages/tbd

Status Category Percentage Covered / Total
🔵 Lines 33.91% 2877 / 8483
🔵 Statements 33.83% 2978 / 8801
🔵 Functions 38.99% 452 / 1159
🔵 Branches 29.98% 1362 / 4543
File CoverageNo changed files found.
Generated in workflow #1013 for commit 531806f by the Vitest Coverage Report Action

claude added 3 commits June 13, 2026 18:47
- Lead with a concrete value prop: a drop-in meta-skill that helps any
  coding agent ship higher-quality, slop-free code
- Broaden the problem statement beyond memory loss to recurring poor
  engineering and process drift that degrade a repo over time
- Replace the generic "Knowledge/Processes/Tasks (what to know...)"
  pillar labels with concrete ones: Engineering knowledge (naming the
  tech covered: TypeScript, Python, TDD, golden testing, Convex),
  Spec-driven processes, Disciplined task tracking
- Note that guidelines are easy to add and, once in tbd, get applied
  systematically rather than pasted into prompts

https://claude.ai/code/session_01RDZ8fshcw3u6tqQywZBTPB
- Tagline: "helps you and your coding agents" (not just the agent)
- Rewrite the usage paragraph to convey that tbd is strictly additive:
  it routes agents to extra knowledge and task tracking when relevant
  without overly constraining them

https://claude.ai/code/session_01RDZ8fshcw3u6tqQywZBTPB
- Rename the processes pillar to "Specs, planning, and code review
  processes" to name what it covers
- Credit the idea as borrowed from Beads, and note tbd beads are plain
  Markdown files with YAML frontmatter, more transparent and less
  glitchy than bd's SQLite database

https://claude.ai/code/session_01RDZ8fshcw3u6tqQywZBTPB
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants