Repair documentation capability parity

[jonathanong]

What changed

• index all stable CLI command docs, including data-pw, effects, rsc-callers, and registry-extension
• document the complete runtime Node export inventory and omitted query APIs
• update the no-mistakes skill for resource edges, baseUrl resolution, and analyzeProject reuse
• narrow the README capability claim and replace hardcoded docs coverage lists with source-derived CLI/export guards

Why

The implementation had stable commands and Node exports that agents could not discover from the primary indexes. The old docs test passed while omissions remained because its inventory was hand-maintained.

Checks

• CARGO_TARGET_DIR=/private/tmp/docs-target RUSTC_WRAPPER= cargo test -p no-mistakes --test docs_coverage
• node --test packages/no-mistakes/scripts/type-docs.test.js packages/no-mistakes/scripts/api.test.js
• cargo fmt --all -- --check
• git diff --check
• GitHub Test CI: passed (Rust coverage, lint/package, JavaScript, AST, benchmarks)

Root cause

CLI and API documentation inventories drifted from source; the skill also retained a stale baseUrl limitation.

Shepherd Journal

• Codecov's transient publishing state cleared; the final Rust and JavaScript coverage gates passed.
• CodSpeed reported a 10.21% memory delta in facts/imports_and_symbols. This PR changes only documentation and a documentation test, with no production analysis or benchmark code path changes, so the result is treated as measurement noise rather than a code regression.
• Fixed the nested Clap command documentation guard and corrected the JSONC example; the CodSpeed memory-only regression is unrelated to these docs/test changes and was left for explicit acknowledgement rather than a speculative performance change.
• Codecov currently reports no coverage information on head while the Rust/coverage workflow remains in progress; this is an external upload-state failure, so no source change is warranted.
• Codecov currently reports no coverage information on head while Rust coverage is still in progress; the external upload-state failure does not warrant a source change.

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@skills/no-mistakes/references/limits-and-fallbacks.md`:
- Around line 12-16: Update the compilerOptions example near the baseUrl
explanation so its fenced language is jsonc, or move the // explanation outside
the code block; keep the tsconfig.json example and its meaning unchanged.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 19e678f2-306d-44ea-916b-f76f5ab46f9b

📥 Commits

Reviewing files that changed from the base of the PR and between 287453d and f96e881.

📒 Files selected for processing (6)

• README.md
• crates/no-mistakes/tests/docs_coverage.rs
• docs/cli/README.md
• docs/node-api.md
• skills/no-mistakes/SKILL.md
• skills/no-mistakes/references/limits-and-fallbacks.md

Fixed nested Clap command documentation coverage and corrected the JSONC example; CodSpeed's docs-only memory result remains explicitly documented for acknowledgement.

[coderabbitai]

Actionable comments posted: 3

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@crates/no-mistakes/tests/docs_coverage.rs`:
- Around line 193-209: Update the runtime inventory parsing in the documentation
coverage test around documented_exports to extract both the export name and API
column, then assert the API value is non-empty for every documented export.
Preserve the exact export-set comparison while strengthening the per-row
validation associated with source_exports.

In `@crates/no-mistakes/tests/support/docs_coverage_cli_helpers.rs`:
- Around line 4-16: Replace the raw recursive traversal in rust_sources with
Git-aware source discovery that excludes ignored files and safely handles
directory symlinks; preferably pass a Git-visible file inventory into the test,
otherwise use an ignore-aware walker while preserving sorted Rust-file results.
- Around line 21-50: Update subcommand_enums to parse command attributes
structurally instead of searching for the exact #[command(subcommand)] byte
sequence, including multiline attributes with whitespace around subcommand. Add
a fixture covering a multiline #[command(...)] attribute and ensure the existing
parent struct and command field extraction remains correct.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: a89d0456-e875-4fc2-b227-5958eb119e3d

📥 Commits

Reviewing files that changed from the base of the PR and between f96e881 and 7cc1def.

📒 Files selected for processing (3)

• crates/no-mistakes/tests/docs_coverage.rs
• crates/no-mistakes/tests/support/docs_coverage_cli_helpers.rs
• skills/no-mistakes/references/limits-and-fallbacks.md

🚧 Files skipped from review as they are similar to previous changes (1)

• skills/no-mistakes/references/limits-and-fallbacks.md

Hardened runtime API-row validation, switched command discovery to Git-aware files with structural syn parsing, and added a multiline attribute fixture.

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@crates/no-mistakes/tests/support/docs_coverage_cli_helpers.rs`:
- Around line 64-96: The subcommand_enums function only scans top-level items,
so Clap structs inside inline modules are omitted. Update its item traversal to
recursively inspect each Item::Mod content while preserving the existing struct
and field extraction, and add a fixture under fixtures/docs-coverage/ covering
an inline subcommand declaration.
- Around line 82-92: The clap subcommand type extraction must unwrap a
single-segment Option before enum lookup. Update the Type::Path handling to
inspect Option’s GenericArgument::Type and return the inner type’s enum name,
while preserving existing named-path behavior; add regression coverage for an
optional subcommand field.

🪄 Autofix

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: bc6d0eb8-88ec-4f93-8e3a-a984dbff366f

📥 Commits

Reviewing files that changed from the base of the PR and between 7cc1def and 5d13b3f.

📒 Files selected for processing (3)

• crates/no-mistakes/tests/docs_coverage.rs
• crates/no-mistakes/tests/support/docs_coverage_cli_helpers.rs
• fixtures/docs-coverage/multiline-subcommand/fixture.rs

🚧 Files skipped from review as they are similar to previous changes (1)

• crates/no-mistakes/tests/docs_coverage.rs

Added recursive inline-module and optional-subcommand parsing with fixture-backed tests.