Skip to content

docs(witan): add Concepts and Walkthroughs sections to the docs site - #289

Open
blarghmatey wants to merge 3 commits into
mainfrom
witan-docs-concepts-walkthroughs
Open

docs(witan): add Concepts and Walkthroughs sections to the docs site#289
blarghmatey wants to merge 3 commits into
mainfrom
witan-docs-concepts-walkthroughs

Conversation

@blarghmatey

Copy link
Copy Markdown
Member

What are the relevant tickets?

N/A

Description (What does it do?)

Adds two new top-level nav sections to the witan-context docs site, between "Get started" and "Guides":

  • Concepts (docs/concepts/) — the mental model the Guides/Reference/Explanation pages assume: memory.md (what a memory is, the four kinds, the edge types), graph.md (the task/project graph — hierarchy, dependency, and provenance edges, with a worked multi-step example), interfaces.md (CLI vs. direct agent tool calls vs. packaged skills, and when to reach for each).
  • Walkthroughs (docs/walkthroughs/) — one running scenario (the retry-bug task already used in the "Tasks and projects" tutorial) worked through three times, once per interface: cli-driven.md, agent-driven.md, skills-driven.md, plus an index comparing what actually differs between them.

Also wires both sections into zensical.toml nav and cross-links them from docs/index.md, docs/explanation/index.md, and the existing tip box in docs/getting-started/tasks-and-projects.md. No generated or mirrored pages touched.

How can this be tested?

  • uvx zensical@0.0.56 build — reports "No issues found" (this is what CI's link/anchor checker in .github/workflows/docs.yml runs).
  • ./bin/gen_docs.py --check — confirms nothing generated/mirrored drifted.
  • Read through docs/concepts/ and docs/walkthroughs/ for tone/accuracy against the existing Explanation and Guides pages, and check the new nav entries render in the right order (zensical serve locally, or the RTD preview once CI runs on this PR).

Guides and Explanation assume readers already have a mental model of
memory kinds, the task/project graph, and the CLI/agent/skills split.
Concepts gives that model directly; Walkthroughs works one scenario
through all three interfaces so the difference between them is
concrete rather than asserted.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012wCHjz7AaaqLeQirfPYaXD
Copilot AI balanced review requested due to automatic review settings August 25, 2026 21:23

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Adds Concepts and Walkthroughs sections to explain witan’s mental model and demonstrate its CLI, agent, and skill interfaces.

Changes:

  • Documents memory, graph, and interface concepts.
  • Adds three parallel workflow walkthroughs.
  • Updates navigation and cross-links.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 12 comments.

Show a summary per file
File Description
zensical.toml Registers the new navigation sections.
docs/concepts/index.md Introduces the Concepts section.
docs/concepts/memory.md Explains memory kinds and relationships.
docs/concepts/graph.md Describes task and project graphs.
docs/concepts/interfaces.md Compares CLI, agent, and skill interfaces.
docs/walkthroughs/index.md Introduces and compares walkthroughs.
docs/walkthroughs/cli-driven.md Demonstrates the CLI workflow.
docs/walkthroughs/agent-driven.md Demonstrates direct MCP tool usage.
docs/walkthroughs/skills-driven.md Demonstrates skill-guided task handling.
docs/index.md Links the new sections from the homepage.
docs/getting-started/tasks-and-projects.md Cross-links the new material.
docs/explanation/index.md Directs conceptual readers appropriately.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Comment thread docs/walkthroughs/agent-driven.md
Comment thread docs/walkthroughs/skills-driven.md
Comment thread docs/concepts/graph.md Outdated
Comment thread docs/concepts/memory.md Outdated
Comment thread docs/concepts/memory.md Outdated
Comment thread docs/walkthroughs/agent-driven.md Outdated
Comment thread docs/concepts/interfaces.md
Comment thread docs/walkthroughs/cli-driven.md Outdated
Comment thread docs/walkthroughs/skills-driven.md
Comment thread docs/concepts/graph.md Outdated
blarghmatey and others added 2 commits August 25, 2026 19:51
Agents consuming the docs site shouldn't have to parse rendered HTML.
bin/gen_llms.py derives llms.txt (llmstxt.org), llms-full.txt, and a
per-page markdown mirror from docs/ and zensical.toml's nav after the
HTML build, so there's no second copy of the site structure to hand-
maintain. Runs as plain python3, not uv run, so it doesn't drag the
uv workspace into the RTD build the way the rest of bin/ deliberately
does not.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012wCHjz7AaaqLeQirfPYaXD
- task_create() examples (agent-driven, skills-driven, concepts/graph)
  omitted the required `description` argument.
- concepts/graph.md's DiscoveredFrom diagram had the edge reversed:
  task_link/task_create write it new-task -> source-task, not the
  other way round.
- concepts/memory.md: `memory_for_contract` needs `kind` to run the
  code-graph lookup at all; `kind` doesn't gate what `recall` returns
  by default (it searches every kind unless the caller filters);
  `agent_context` has no automatic scope or expiry, contra the prior
  wording; `related_to` was missing from the edge table.
- walkthroughs/agent-driven.md overstated what `recall` expands
  across (applies_to/related_to + siblings, not "any linked memory").
- walkthroughs/cli-driven.md showed a claim output format and a
  --ready listing that don't match witan/cli/tasks.py.
- walkthroughs/skills-driven.md closed the task without storing the
  lesson the overview page claims all three walkthroughs preserve;
  added the memory_store step and corrected the comparison table.
- concepts/interfaces.md: missing connector made a sentence unparseable.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012wCHjz7AaaqLeQirfPYaXD
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