/pr-review-fix
Fix specific issues from a PR review using feature-dev
> /plugin marketplace add LeanAndMean/mach10 > /plugin install mach10@LeanAndMean-mach10
How it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/pr-review-fix
Context preview
What this command does when you run it.
Fix specific issues from a PR review using feature-dev
Command definition
pr-review-fix.mddescription: Fix specific issues from a PR review using feature-dev
argument-hint: <pr-number> [--review-comment <id>] [--assessment-comment <id>] [findings] [context]
allowed-tools: Bash, Read, Grep, Glob, Task, TaskCreate, TaskUpdate, Edit, Write, Skill, AskUserQuestion
Fix Review Issues
You are fixing specific issues identified in a PR review. This command gathers context from the PR and its review comments, then delegates to the feature-dev workflow for implementation.
**User input:** $ARGUMENTS
Step 0: Parse input and create task list
The user's input typically contains:
- A **PR number** (required)
- **`--review-comment <id>`** flag with a numeric comment ID (optional)
- **`--assessment-comment <id>`** flag with a numeric comment ID (optional)
- **Finding identifiers to fix** -- space-separated F/S identifiers (e.g., `F1 F2 S3`) or bare numbers as fallback (optional)
- Additional context or constraints (optional)
Example inputs:
- `108` (PR only -- read the PR and determine which review findings to fix)
- `108 F1 F2 S3`
- `108 --review-comment 1234567890 --assessment-comment 1234567891 F1 F2 S3`
- `108 --review-comment 1234567890 F1 S2 focus on error handling`
Extract the PR number. Parse `--review-comment` and `--assessment-comment` flags if present (each followed by a numeric ID). If finding identifiers are provided (F/S prefixed or bare numbers), note them. If the input is ambiguous, ask the user to clarify.
After parsing input, create the progress-tracking task list. Create a task for Step 0 and immediately mark it in progress. Then create tasks for each of the remaining 4 steps one at a time, in step order, all starting as pending. Task list display order matches creation order, so each task must be a separate sequential call -- do not batch multiple task creations in a single message. Store each returned task ID for later use -- do not assume IDs are sequential.
| Task | Subject | activeForm | |------|---------|------------| | Step 0 | Step 0: Parse input and create task list | Parsing input | | Step 1 | Step 1: Gather PR and review context | Gathering context | | Step 2 | Step 2: Identify issues to fix | Identifying issues | | Step 3 | Step 3: Delegate review fixes to feature-dev | Delegating review fixes to feature-dev | | Step 4 | Step 4: Recommend next steps | Recommending next steps |
Mark Step 0 complete.
Step 1: Gather PR and review context
Mark Step 1 in progress.
Read the PR title and description:
gh pr view <pr-number>
Locate the review comment
**If `--review-comment` was provided:** Fetch the specific comment by ID:
gh api repos/:owner/:repo/issues/comments/<review-comment-id>
Extract the `body` field from the JSON response. This is the review comment content.
**If `--review-comment` was NOT provided (fallback):** Fetch all comments as JSON and find the review heuristically:
gh pr view <pr-number> --json comments
Parse the JSON array and search from the END (most recent first) for the first comment whose body contains the HTML marker `<!-- mach10-review -->`. If no comment contains the marker, fall back to finding the last comment with the structured review format (Critical/Important/Suggestions sections and model attribution).
Locate the assessment comment (optional)
**If `--assessment-comment` was provided:** Fetch it by ID:
gh api repos/:owner/:repo/issues/comments/<assessment-comment-id>
**If not provided:** This is optional context. Do not attempt to locate the assessment heuristically — proceed without it.
Save the review comment content for use in Step 3.
Mark Step 1 complete.
Step 2: Identify issues to fix
Mark Step 2 in progress.
**If finding identifiers were provided in the input:**
- Match identifiers against the F/S labels in the review comment (e.g., `F1` matches `**F1:**`). If bare numbers were given, match by sequential position across Critical and Important sections. If the review comment lacks F/S labels (e.g., older reviews), fall back to matching by ordinal position within each severity section. Extract the full finding descriptions.
**If only a PR number was provided with no specific finding identifiers:**
- Present all review findings to the user, organized by severity.
- Recommend which to fix in this session using batch sizing heuristics:
- **Simple one-line fixes:** up to ~10 at once
- **Moderate fixes:** ~6 at a time
- **Deep or complex fixes:** no more than ~3 at a time
- Group similar issues together
- Use `AskUserQuestion` with `multiSelect: true` to let the user select which issues to fix. If there are 4 or fewer findings, list each as a separate option. If there are more than 4 findings, group them by severity as options (e.g., "All critical findings (3)", "All important findings (5)"). The built-in "Other" option allows the user to specify individual finding identifiers (e.g., F1 S3) if they prefer a custom selection.
Mark Step 2 complete.
Step 3: Delegate review fixes to feature-dev
Mark Step 3 in progress.
Mark Step 3 complete when all sub-tasks of the delegation below are completed.
Use the Skill tool to invoke `/feature-dev:feature-dev` with the following context:
> Read PR #<pr-number> (title and description only -- do not fetch all comments). The review comment content is provided below. Implement fixes for the identified issues using the 7-phase development plan. IMPORTANT: Create a task for each phase of the 7-phase development plan before starting work. Use `"Step 3.M: <action>"` as the subject format (e.g., `"Step 3.1: Understand the codebase"`). Mark each phase in progress when starting and complete when finishing. Do not skip any phases. > > **Findings to fix (from Step 2):** <list the resolved finding identifiers and their one-line descriptions> > > Fix only the findings listed above. Do not fix other findings in the review comment. > > **Review comment:** > <paste the full review comment content retrieved in Step 1>
Read more
description: Fix specific issues from a PR review using feature-dev argument-hint: <pr-number> [--review-comment <id>] [--assessment-comment <id>] [findings] [context] allowed-tools: Bash, Read, Grep, Glob, Task, TaskCreate, TaskUpdate, Edit, Write, Skill, AskUserQuestion
Fix Review Issues
You are fixing specific issues identified in a PR review. This command gathers context from the PR and its review comments, then delegates to the feature-dev workflow for implementation.
**User input:** $ARGUMENTS
Step 0: Parse input and create task list
The user's input typically contains:
- A **PR number** (required)
- **`--review-comment <id>`** flag with a numeric comment ID (optional)
- **`--assessment-comment <id>`** flag with a numeric comment ID (optional)
- **Finding identifiers to fix** -- space-separated F/S identifiers (e.g., `F1 F2 S3`) or bare numbers as fallback (optional)
- Additional context or constraints (optional)
Example inputs:
- `108` (PR only -- read the PR and determine which review findings to fix)
- `108 F1 F2 S3`
- `108 --review-comment 1234567890 --assessment-comment 1234567891 F1 F2 S3`
- `108 --review-comment 1234567890 F1 S2 focus on error handling`
Extract the PR number. Parse `--review-comment` and `--assessment-comment` flags if present (each followed by a numeric ID). If finding identifiers are provided (F/S prefixed or bare numbers), note them. If the input is ambiguous, ask the user to clarify.
After parsing input, create the progress-tracking task list. Create a task for Step 0 and immediately mark it in progress. Then create tasks for each of the remaining 4 steps one at a time, in step order, all starting as pending. Task list display order matches creation order, so each task must be a separate sequential call -- do not batch multiple task creations in a single message. Store each returned task ID for later use -- do not assume IDs are sequential.
| Task | Subject | activeForm | |------|---------|------------| | Step 0 | Step 0: Parse input and create task list | Parsing input | | Step 1 | Step 1: Gather PR and review context | Gathering context | | Step 2 | Step 2: Identify issues to fix | Identifying issues | | Step 3 | Step 3: Delegate review fixes to feature-dev | Delegating review fixes to feature-dev | | Step 4 | Step 4: Recommend next steps | Recommending next steps |
Mark Step 0 complete.
Step 1: Gather PR and review context
Mark Step 1 in progress.
Read the PR title and description:
gh pr view <pr-number>
Locate the review comment
**If `--review-comment` was provided:** Fetch the specific comment by ID:
gh api repos/:owner/:repo/issues/comments/<review-comment-id>
Extract the `body` field from the JSON response. This is the review comment content.
**If `--review-comment` was NOT provided (fallback):** Fetch all comments as JSON and find the review heuristically:
gh pr view <pr-number> --json comments
Parse the JSON array and search from the END (most recent first) for the first comment whose body contains the HTML marker `<!-- mach10-review -->`. If no comment contains the marker, fall back to finding the last comment with the structured review format (Critical/Important/Suggestions sections and model attribution).
Locate the assessment comment (optional)
**If `--assessment-comment` was provided:** Fetch it by ID:
gh api repos/:owner/:repo/issues/comments/<assessment-comment-id>
**If not provided:** This is optional context. Do not attempt to locate the assessment heuristically — proceed without it.
Save the review comment content for use in Step 3.
Mark Step 1 complete.
Step 2: Identify issues to fix
Mark Step 2 in progress.
**If finding identifiers were provided in the input:**
- Match identifiers against the F/S labels in the review comment (e.g., `F1` matches `**F1:**`). If bare numbers were given, match by sequential position across Critical and Important sections. If the review comment lacks F/S labels (e.g., older reviews), fall back to matching by ordinal position within each severity section. Extract the full finding descriptions.
**If only a PR number was provided with no specific finding identifiers:**
- Present all review findings to the user, organized by severity.
- Recommend which to fix in this session using batch sizing heuristics:
- **Simple one-line fixes:** up to ~10 at once
- **Moderate fixes:** ~6 at a time
- **Deep or complex fixes:** no more than ~3 at a time
- Group similar issues together
- Use `AskUserQuestion` with `multiSelect: true` to let the user select which issues to fix. If there are 4 or fewer findings, list each as a separate option. If there are more than 4 findings, group them by severity as options (e.g., "All critical findings (3)", "All important findings (5)"). The built-in "Other" option allows the user to specify individual finding identifiers (e.g., F1 S3) if they prefer a custom selection.
Mark Step 2 complete.
Step 3: Delegate review fixes to feature-dev
Mark Step 3 in progress.
Mark Step 3 complete when all sub-tasks of the delegation below are completed.
Use the Skill tool to invoke `/feature-dev:feature-dev` with the following context:
> Read PR #<pr-number> (title and description only -- do not fetch all comments). The review comment content is provided below. Implement fixes for the identified issues using the 7-phase development plan. IMPORTANT: Create a task for each phase of the 7-phase development plan before starting work. Use `"Step 3.M: <action>"` as the subject format (e.g., `"Step 3.1: Understand the codebase"`). Mark each phase in progress when starting and complete when finishing. Do not skip any phases. > > **Findings to fix (from Step 2):** <list the resolved finding identifiers and their one-line descriptions> > > Fix only the findings listed above. Do not fix other findings in the review comment. > > **Review comment:** > <paste the full review comment content retrieved in Step 1>
A development methodology for agentic coding -- and a Claude Code plugin that implements it.
Other commands on mach10.
issue-assessment
Read a GitHub issue, perform an independent assessment, and present findings
issue-create
Create a structured GitHub issue from current context or description
issue-implement
Implement a specific stage of an issue's implementation plan using feature-dev
issue-plan-review
Read a GitHub issue and all comments, review the implementation plan, independently assess each finding, and present findings
issue-plan
Read a GitHub issue, analyze the codebase, and create a staged implementation plan

