Open
Conversation
Contributor
📝 WalkthroughWalkthroughDocumentation is updated for runbook format: Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 7 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (7 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
⚔️ Resolve merge conflicts
Review rate limit: 4/5 reviews remaining, refill in 12 minutes. Comment |
Deploying rundown with
|
| Latest commit: |
99dec78
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://c6907342.rundown-7hl.pages.dev |
| Branch Preview URL: | https://2026-04-30-cleanup-ii.rundown-7hl.pages.dev |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@docs/FORMAT.md`:
- Around line 59-63: The inline grammar currently disallows spaces around commas
and inside brackets; update inline_sequence and inline_output_sequence to accept
optional whitespace around separators and at the bracket boundaries.
Specifically, change inline_sequence (symbol inline_sequence) to allow optional
ws before/after each comma and optional ws after the opening "[" and before the
closing "]", and change inline_output_sequence (symbol inline_output_sequence)
similarly so the pattern becomes "[" ws? variable_name ( ws? "," ws?
variable_name )* ws? "]" and "[" ws? ( output_sequence_entry ( ws? "," ws?
output_sequence_entry )* )? ws? "]" respectively, leaving output_sequence_entry,
output_entry, and quoted_output_entry unchanged.
🪄 Autofix (Beta)
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: Path: .coderabbit.yaml
Review profile: ASSERTIVE
Plan: Pro
Run ID: e0ab62ba-4ef7-4694-97da-9e029a7f338c
📒 Files selected for processing (2)
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📜 Review details
⏰ Context from checks skipped due to timeout of 900000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (2)
- GitHub Check: performance-test
- GitHub Check: Cloudflare Pages
🧰 Additional context used
📓 Path-based instructions (2)
docs/**/*.md
⚙️ CodeRabbit configuration file
docs/**/*.md: Documentation changes. Focus on:
- Accuracy against current CLI behavior and public Rundown specification.
- Updating linked docs together when public syntax, output schema, policy behavior, or runbook discovery changes.
- Keeping examples executable and aligned with package scripts.
Files:
docs/FORMAT.md
packages/claude-code-plugin/skills/**/*.md
⚙️ CodeRabbit configuration file
packages/claude-code-plugin/skills/**/*.md: Plugin skill instructions. Focus on:
- Accuracy of tool names, command examples, and workflow sequencing.
- No stale references to removed runbooks, templates, or package commands.
- Clear handoff instructions when skills generate plans or delegate work.
Files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
🧠 Learnings (18)
📓 Common learnings
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/write-plan.runbook.md:68-91
Timestamp: 2026-04-06T02:43:13.971Z
Learning: In tobyhede/rundown, `write-plan.runbook.md` and `review-plan.runbook.md` under `packages/claude-code-plugin/runbooks/planning/` are intentionally separate runbooks. `write-plan` creates the plan; `review-plan` validates it. They are composed at a higher orchestration level (e.g., `meta/end-to-end-test.runbook.md` or manual invocation with `--var PlanPath=...`). Do NOT flag the absence of a direct chain between these two runbooks as a workflow gap — coupling them would reduce flexibility by design.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review-plan.runbook.md:7-28
Timestamp: 2026-04-06T02:43:16.222Z
Learning: In tobyhede/rundown, runbook template variables (e.g., `{{ PlanPath }}`, `{{ TargetRunbook }}`) that are not supplied at invocation time are preserved as literal text per spec. The first step that reads the unresolved variable will fail with a clear error, so adding explicit upfront validation guards (e.g., `test -n "{{ PlanPath }}"`) is considered unnecessary boilerplate without meaningful UX improvement. Do not flag missing validation guards for required runbook variables in future reviews.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 202
File: packages/cli/src/commands/status.ts:22-23
Timestamp: 2026-04-15T05:28:44.644Z
Learning: In tobyhede/rundown, the removal of `--json` from all CLI commands (status, ls, pass, fail, run, etc.) in PR `#202` is an intentional breaking change. JSON output is now the default for all commands; `--json` is fully removed and must NOT be re-added as a deprecated alias. Scripts that previously passed `--json` simply need to remove the flag — they continue to receive JSON output. This is documented and tested in `packages/cli/__tests__/commands/json-output.test.ts`. Do not flag the removal of `--json` as a backward-compatibility concern in future reviews.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review-plan.runbook.md:36-40
Timestamp: 2026-04-06T03:16:22.299Z
Learning: In tobyhede/rundown (`packages/cli/src/helpers/resolve-runbook.ts`), `resolveRunbookFile` performs recursive subdirectory scanning within each source root (project at `.claude/rundown/runbooks/`, plugin at `$CLAUDE_PLUGIN_ROOT/runbooks/`, and bundled). A bare filename like `review-technical-accuracy.runbook.md` or `retry-success.runbook.md` will resolve correctly even when the actual file lives in a subdirectory (e.g., `review/` or `retries/`). Do NOT flag bare runbook filename references in runbook delegate lists as broken paths requiring a subdirectory prefix — name-based discovery handles this transparently.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 210
File: packages/claude-code-plugin/runbooks/meta/end-to-end-test.runbook.md:27-31
Timestamp: 2026-04-27T06:40:40.803Z
Learning: In tobyhede/rundown `*.runbook.md` files, a runbook-list entry (e.g., `- planning/review-plan.runbook.md`) in a step WITHOUT a `- DELEGATE` directive is an intentional inline linkage pattern: the parent runbook walks the referenced runbook in-session as an implicit substep (per SPEC §3.2). This is distinct from a step WITH `- DELEGATE`, which dispatches an out-of-process subagent. Both are valid runbook-list patterns with different execution boundaries. Do NOT flag the absence of `- DELEGATE` on a runbook-list step as a missing directive — it is a deliberate choice between in-session (inline) and out-of-process (delegate) execution.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 92
File: runbooks/README.md:49-49
Timestamp: 2026-03-10T12:24:15.826Z
Learning: In the `tobyhede/rundown` repository, the requirement that every scenario must include at least one of a top-level `result:` or an `expect.result` field is authoritatively documented at `docs/SCENARIOS.md` line 46. The `runbooks/README.md` intentionally keeps only a brief usage hint and defers to `docs/SCENARIOS.md` for full schema details.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 187
File: packages/claude-code-plugin/runbooks/planning/review/review-technical-accuracy.runbook.md:29-31
Timestamp: 2026-04-07T22:54:01.013Z
Learning: In tobyhede/rundown, `PASS COMPLETE` on the "Review the plan for [dimension]" step (Step 3) in review sub-runbooks under `packages/claude-code-plugin/runbooks/planning/review/` (e.g., `review-technical-accuracy.runbook.md`, `review-build-runtime.runbook.md`) is intentional. When a review passes with no findings, there is nothing to emit and Step 4 (write the review artifact) is intentionally skipped. The synthesize runbook (`review-synthesize.runbook.md`) handles the absence of per-dimension `plan-review-{{ RunId }}.json` artifacts correctly. Do NOT flag `PASS COMPLETE` on the review step as a control-flow bug that skips artifact emission.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/write-plan.runbook.md:94-97
Timestamp: 2026-04-06T04:20:08.732Z
Learning: In tobyhede/rundown `.runbook.md` files, a blank line between a step heading (e.g., `## 10. Verify Plan Structure`) and its transition directives (e.g., `- PASS COMPLETE`, `- FAIL GOTO 8`) is cosmetic only. The Rundown parser correctly parses the control-flow regardless of the blank line. Do NOT flag blank lines between headings and transition directives as a potential parsing or control-flow issue — `rd check` validates steps correctly and scenario tests confirm the expected routing behavior.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/meta/end-to-end-test.runbook.md:8-18
Timestamp: 2026-04-06T02:43:17.509Z
Learning: In `tobyhede/rundown` runbooks (e.g., `packages/claude-code-plugin/runbooks/meta/end-to-end-test.runbook.md`), empty or unset `vars:` values are intentionally left without shell-level validation guards. By design, undefined/empty template variables are preserved as literal `{{ VariableName }}` text, which produces a clear error when the step attempts to resolve the runbook reference or command. Adding input-guard steps to every runbook that accepts variables is considered unnecessary boilerplate; the `vars:` declaration and skill documentation are the intended contract.
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review/review-structural-integrity.runbook.md:1-22
Timestamp: 2026-04-06T02:43:03.967Z
Learning: In tobyhede/rundown, the rundown variable priority system has frontmatter `vars:` at priority 4 and inherited delegation variables at priority 5. Because priority 4 overrides priority 5, child runbooks that receive variables via parent delegation (e.g., `PlanPath` passed from `review-plan.runbook.md`) must NOT declare those variables in their own `vars:` frontmatter block — doing so would shadow/override the delegated value with the default. Omitting `vars:` declarations for inherited delegation variables in child runbooks is intentional and correct. Do not flag missing `vars:` declarations for inherited variables as a documentation gap in future reviews of files under `packages/claude-code-plugin/runbooks/planning/review/`.
📚 Learning: 2026-04-06T03:16:01.196Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review/review-structural-integrity.runbook.md:18-18
Timestamp: 2026-04-06T03:16:01.196Z
Learning: In `packages/claude-code-plugin/runbooks/planning/review/` runbooks (e.g., `review-structural-integrity.runbook.md`), substep headings use intentionally terse noun-phrase identifiers (e.g., "Step ordering logical", "No circular dependencies", "Clear completion criteria") rather than full grammatical sentences. Do not flag these as grammar issues — this is a deliberate, consistent style across all sibling substep headings in these review runbooks.
Applied to files:
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-27T06:40:40.803Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 210
File: packages/claude-code-plugin/runbooks/meta/end-to-end-test.runbook.md:27-31
Timestamp: 2026-04-27T06:40:40.803Z
Learning: In tobyhede/rundown `*.runbook.md` files, a runbook-list entry (e.g., `- planning/review-plan.runbook.md`) in a step WITHOUT a `- DELEGATE` directive is an intentional inline linkage pattern: the parent runbook walks the referenced runbook in-session as an implicit substep (per SPEC §3.2). This is distinct from a step WITH `- DELEGATE`, which dispatches an out-of-process subagent. Both are valid runbook-list patterns with different execution boundaries. Do NOT flag the absence of `- DELEGATE` on a runbook-list step as a missing directive — it is a deliberate choice between in-session (inline) and out-of-process (delegate) execution.
Applied to files:
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T03:16:01.394Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review/review-technical-accuracy.runbook.md:48-54
Timestamp: 2026-04-06T03:16:01.394Z
Learning: In tobyhede/rundown runbook markdown files (e.g., `*.runbook.md` under `packages/claude-code-plugin/runbooks/`), the final step of a runbook does NOT require an explicit execution directive (such as `- DEFER`) because it relies on default terminal behavior. Do not flag missing `- DEFER` or transition markers on a runbook's last step — it is intentionally omitted and executes correctly via implicit terminal semantics.
Applied to files:
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T04:20:08.732Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/write-plan.runbook.md:94-97
Timestamp: 2026-04-06T04:20:08.732Z
Learning: In tobyhede/rundown `.runbook.md` files, a blank line between a step heading (e.g., `## 10. Verify Plan Structure`) and its transition directives (e.g., `- PASS COMPLETE`, `- FAIL GOTO 8`) is cosmetic only. The Rundown parser correctly parses the control-flow regardless of the blank line. Do NOT flag blank lines between headings and transition directives as a potential parsing or control-flow issue — `rd check` validates steps correctly and scenario tests confirm the expected routing behavior.
Applied to files:
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T03:16:03.975Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/write-plan.runbook.md:11-11
Timestamp: 2026-04-06T03:16:03.975Z
Learning: In tobyhede/rundown `.runbook.md` files, the list items immediately following a step heading (e.g., `- PASS CONTINUE`, `- FAIL STOP`, `- FAIL GOTO N`, `- PASS COMPLETE`) are Rundown spec transition/control-flow directives that must appear directly after the heading with no blank line. Do NOT flag these as markdownlint MD022 (blanks-around-headings) violations — they are not standard Markdown prose lists and the blank-line requirement does not apply to them in runbook files.
Applied to files:
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-03-15T08:52:14.181Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 109
File: docs/FORMAT.md:19-19
Timestamp: 2026-03-15T08:52:14.181Z
Learning: Document that the ws lexical rule is a general definition (space or tab) while individual productions may restrict to spaces only. When reviewing docs and related code, ensure this nuance is clearly stated (e.g., in docs/FORMAT.md and related code comments), and reference the relevant code paths and tests (e.g., parser helpers and tests around PASS usage and tab handling). Prefer explicit notes over implying universal tab support, and consider adding automated checks or CI validation to enforce consistency between the lexical rule and production-specific behavior.
Applied to files:
docs/FORMAT.md
📚 Learning: 2026-03-31T01:30:36.008Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 166
File: packages/claude-code-plugin/skills/writing-runbooks/SKILL.md:140-140
Timestamp: 2026-03-31T01:30:36.008Z
Learning: In this repository, when writing Handlebars templates in Markdown, spaced and unspaced variable forms are both valid. Do not treat `{{ item }}` (with whitespace around the variable name) as a style violation; it is intentionally supported and is used in documentation for readability. Only flag Handlebars syntax issues if they are syntactically invalid beyond whitespace (e.g., malformed braces or invalid expression).
Applied to files:
docs/FORMAT.mdpackages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T02:43:03.967Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review/review-structural-integrity.runbook.md:1-22
Timestamp: 2026-04-06T02:43:03.967Z
Learning: In tobyhede/rundown, the rundown variable priority system has frontmatter `vars:` at priority 4 and inherited delegation variables at priority 5. Because priority 4 overrides priority 5, child runbooks that receive variables via parent delegation (e.g., `PlanPath` passed from `review-plan.runbook.md`) must NOT declare those variables in their own `vars:` frontmatter block — doing so would shadow/override the delegated value with the default. Omitting `vars:` declarations for inherited delegation variables in child runbooks is intentional and correct. Do not flag missing `vars:` declarations for inherited variables as a documentation gap in future reviews of files under `packages/claude-code-plugin/runbooks/planning/review/`.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T02:42:58.905Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review/review-build-runtime.runbook.md:1-22
Timestamp: 2026-04-06T02:42:58.905Z
Learning: In tobyhede/rundown, runbook frontmatter `vars:` declarations carry variable priority 4, which overrides inherited delegation variables at priority 5. Child runbooks that rely on `PlanPath` (or any variable) being passed down from a parent via delegation must NOT declare that variable in their own `vars:` frontmatter — doing so would shadow the parent's value with the default (typically empty string). Omitting the `vars:` entry is the correct pattern to enable delegation inheritance. This applies to all review sub-runbooks under `packages/claude-code-plugin/runbooks/planning/review/`.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T02:43:17.509Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/meta/end-to-end-test.runbook.md:8-18
Timestamp: 2026-04-06T02:43:17.509Z
Learning: In `tobyhede/rundown` runbooks (e.g., `packages/claude-code-plugin/runbooks/meta/end-to-end-test.runbook.md`), empty or unset `vars:` values are intentionally left without shell-level validation guards. By design, undefined/empty template variables are preserved as literal `{{ VariableName }}` text, which produces a clear error when the step attempts to resolve the runbook reference or command. Adding input-guard steps to every runbook that accepts variables is considered unnecessary boilerplate; the `vars:` declaration and skill documentation are the intended contract.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T02:43:16.222Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review-plan.runbook.md:7-28
Timestamp: 2026-04-06T02:43:16.222Z
Learning: In tobyhede/rundown, runbook template variables (e.g., `{{ PlanPath }}`, `{{ TargetRunbook }}`) that are not supplied at invocation time are preserved as literal text per spec. The first step that reads the unresolved variable will fail with a clear error, so adding explicit upfront validation guards (e.g., `test -n "{{ PlanPath }}"`) is considered unnecessary boilerplate without meaningful UX improvement. Do not flag missing validation guards for required runbook variables in future reviews.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T02:43:13.971Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/write-plan.runbook.md:68-91
Timestamp: 2026-04-06T02:43:13.971Z
Learning: In tobyhede/rundown, `write-plan.runbook.md` and `review-plan.runbook.md` under `packages/claude-code-plugin/runbooks/planning/` are intentionally separate runbooks. `write-plan` creates the plan; `review-plan` validates it. They are composed at a higher orchestration level (e.g., `meta/end-to-end-test.runbook.md` or manual invocation with `--var PlanPath=...`). Do NOT flag the absence of a direct chain between these two runbooks as a workflow gap — coupling them would reduce flexibility by design.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-20T07:26:57.555Z
Learnt from: CR
Repo: tobyhede/rundown PR: 0
File: CLAUDE.md:0-0
Timestamp: 2026-04-20T07:26:57.555Z
Learning: Runbooks are discovered from multiple sources with priority: Project (`.rundown/runbooks/`) → Plugin (`$CLAUDE_PLUGIN_ROOT/runbooks/`) → Bundled (CLI package `dist/runbooks/`). Use `namespace:name` syntax for explicit source targeting (e.g., `rundown:write-plan` for plugin namespace).
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T03:16:22.299Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review-plan.runbook.md:36-40
Timestamp: 2026-04-06T03:16:22.299Z
Learning: In tobyhede/rundown (`packages/cli/src/helpers/resolve-runbook.ts`), `resolveRunbookFile` performs recursive subdirectory scanning within each source root (project at `.claude/rundown/runbooks/`, plugin at `$CLAUDE_PLUGIN_ROOT/runbooks/`, and bundled). A bare filename like `review-technical-accuracy.runbook.md` or `retry-success.runbook.md` will resolve correctly even when the actual file lives in a subdirectory (e.g., `review/` or `retries/`). Do NOT flag bare runbook filename references in runbook delegate lists as broken paths requiring a subdirectory prefix — name-based discovery handles this transparently.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T03:16:01.068Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/runbooks/planning/review/review-risk-safety.runbook.md:66-72
Timestamp: 2026-04-06T03:16:01.068Z
Learning: In tobyhede/rundown, the final "Write findings" step in planning review sub-runbooks (e.g., `review-risk-safety.runbook.md`, `review-technical-accuracy.runbook.md`, `review-structural-integrity.runbook.md`, `review-build-runtime.runbook.md` under `packages/claude-code-plugin/runbooks/planning/review/`) intentionally omits a `- DEFER` directive. It is the terminal step and uses the default `PASS COMPLETE / FAIL STOP` semantics to propagate the child runbook's outcome to the parent. Do NOT flag the absence of `- DEFER` on these terminal "Write findings" steps as a missing execution directive — adding `- DEFER` would incorrectly change them from terminating steps to mid-flow deferred steps.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-07T22:53:56.381Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 187
File: packages/claude-code-plugin/runbooks/planning/review/review-build-runtime.runbook.md:29-31
Timestamp: 2026-04-07T22:53:56.381Z
Learning: In tobyhede/rundown, the `PASS COMPLETE` on Step 3 of the review sub-runbooks (`review-build-runtime.runbook.md`, `review-risk-safety.runbook.md`, `review-structural-integrity.runbook.md`, `review-technical-accuracy.runbook.md`) is intentional design. When a review PASS completes (no issues found), no JSON artifact is written — absence of the output file is the implicit "ok" signal. `review-synthesize.runbook.md` is designed to collate only the artifacts that exist, so "ok" reviews that produce no file are correctly and intentionally excluded. Do NOT flag `PASS COMPLETE` before the write step in these runbooks as a control-flow bug.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
📚 Learning: 2026-04-06T02:43:17.379Z
Learnt from: tobyhede
Repo: tobyhede/rundown PR: 177
File: packages/claude-code-plugin/skills/writing-plans/SKILL.md:130-130
Timestamp: 2026-04-06T02:43:17.379Z
Learning: In Rundown `SKILL.md` task definitions for this repo, the `commit` object (including its `files` and `message` sub-keys) is the canonical, schema-compliant representation of staged changes and a commit. The `rdx` tool renders this structured `commit` into `git add`/`git commit` commands at execution time. Therefore, during code review, do not flag an example task definition as missing explicit bash `git add`/`git commit` subtasks when the task uses the `commit` object.
Applied to files:
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md
🔇 Additional comments (2)
docs/FORMAT.md (1)
66-66: Good clarifications on key casing and duplicate OUTPUTS rejection.These statements make parsing behavior and validation constraints explicit and reduce ambiguity for runbook authors.
Also applies to: 107-107
packages/claude-code-plugin/skills/writing-runbooks/SKILL.md (1)
145-151: LGTM — value-source and precedence guidance is clear and consistent.The updated INPUTS/REQUIRED semantics, missing-required behavior, and source-precedence notes are coherent and materially improve authoring correctness.
Also applies to: 240-242, 251-251
Comment on lines
+59
to
+63
| inline_sequence ::= "[" variable_name ( "," variable_name )* "]" | ||
| output_fm_list ::= ( ws "- " quoted_output_entry newline | ||
| | ws "- " output_entry newline )+ | ||
| inline_output_sequence ::= "[" ( output_sequence_entry ( "," output_sequence_entry )* )? "]" | ||
| output_sequence_entry ::= output_entry | quoted_output_entry |
Contributor
There was a problem hiding this comment.
Allow optional whitespace in inline frontmatter sequences.
The new inline grammar currently excludes common YAML inline forms with spaced commas (e.g., [PlanPath, environment]). Please add optional whitespace around separators/items so the grammar matches practical authoring and parser behavior.
Suggested grammar tweak
-inline_sequence ::= "[" variable_name ( "," variable_name )* "]"
+inline_sequence ::= "[" ws? variable_name ( ws? "," ws? variable_name )* ws? "]"
-inline_output_sequence ::= "[" ( output_sequence_entry ( "," output_sequence_entry )* )? "]"
+inline_output_sequence ::= "[" ws? ( output_sequence_entry ( ws? "," ws? output_sequence_entry )* )? ws? "]"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/FORMAT.md` around lines 59 - 63, The inline grammar currently disallows
spaces around commas and inside brackets; update inline_sequence and
inline_output_sequence to accept optional whitespace around separators and at
the bracket boundaries. Specifically, change inline_sequence (symbol
inline_sequence) to allow optional ws before/after each comma and optional ws
after the opening "[" and before the closing "]", and change
inline_output_sequence (symbol inline_output_sequence) similarly so the pattern
becomes "[" ws? variable_name ( ws? "," ws? variable_name )* ws? "]" and "[" ws?
( output_sequence_entry ( ws? "," ws? output_sequence_entry )* )? ws? "]"
respectively, leaving output_sequence_entry, output_entry, and
quoted_output_entry unchanged.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Summary by CodeRabbit
Testing
npm run verifyTargeted tests or checks:
Review Notes
Check all that apply:
docs/SPEC.md,docs/FORMAT.md, andrunbook fixtures were checked.
symlink escape, deny precedence, env leakage, and fail-closed behavior were checked.