Pepe39/.claude/skills/step-08-generate-output

step-08-generate-output

What it does

Compiles all four axis labels into the final JSON output, validates consistency, and generates the deliverable file ready for submission.

Prerequisites

• Steps 04-07 completed (all four axes labeled)

Context

See docs/steps/step10.md for the submission checklist.

Arguments

• id (required): Task ID

Instructions

1. Recover context

Read task_info.md — extract all four labels and their reasoning from the Labels section.

Update progress.md: step 08 status = "in-progress", Started = {timestamp ISO 8601}.

2. Compile labels

Collect from task_info.md:

• Quality: {helpful | unhelpful | wrong}
• Severity: {nit | moderate | critical}
• Context Scope: {diff | file | repo | external}
• Context Array: [{entries}]
• Advanced: {Repo-specific conventions | Context outside changed files | Recent language/library updates | Better implementation approach | False}

3. Validate consistency (GATE)

Run these checks before generating output. If any check fails, report to the user and STOP.

Field validation:

• [ ] Quality is one of: helpful, unhelpful, wrong
• [ ] Severity is one of: nit, moderate, critical
• [ ] Context Scope is one of: diff, file, repo, external
• [ ] Advanced is one of: Repo-specific conventions, Context outside changed files, Recent language/library updates, Better implementation approach, False
• [ ] Context array is valid JSON (array of objects with diff_line, file_path, why)
• [ ] If context_scope is diff, file, or repo, the context array has at least 1 entry
• [ ] If context_scope is external, the context array may be empty

Independence checks (warnings, not blockers):

• If Quality = wrong AND Severity = nit, flag for review: "Verify: the issue the comment tried to flag is truly nit-level, even though the comment is wrong."
• If Quality = helpful AND Advanced = False AND Context Scope = repo, flag for review: "Verify: the comment requires repo-level context but is not considered advanced?"

4. Generate labels.json

Write tasks/{date}/{id}/deliverables/labels.json:

{
  "quality": "{value}",
  "severity": "{value}",
  "context_scope": "{value}",
  "context": [
    {
      "diff_line": "{value_or_null}",
      "file_path": "{value}",
      "why": "{value}"
    }
  ],
  "advanced": "{category_or_False}"
}

Formatting rules:

• Use 2-space indentation
• advanced must be a string matching one of the platform categories or "False"
• diff_line must be a string (e.g., "42") or JSON null when empty. Never use "", "null", or a bare number
• All string values must be properly escaped

5. Generate context.json

Write tasks/{date}/{id}/deliverables/context.json using the table format from table-template.json.

Map the context array entries to the table column IDs:

• _dshks = diff_line value (string or empty string if null)
• ahMYbl = file_path value
• dA0ihr = why value

{
  "rows": [
    {
      "_dshks": "{diff_line_or_empty}",
      "ahMYbl": "{file_path}",
      "dA0ihr": "{why}"
    }
  ]
}

Rules:

• If diff_line is null, use an empty string "" for _dshks
• One row per context array entry
• Use 2-space indentation

6. Generate per-axis deliverables

Write one markdown file per axis inside tasks/{date}/{id}/deliverables/.

IMPORTANT: The ## Reasoning section in each file is the justification the user will paste directly into the annotation platform. It must be self-contained, clear, and ready to copy-paste as-is.

5a. quality.md

# Quality: {id}

- **Comment:** {first 80 chars of body}...
- **File:** {file_path}:{diff_line}
- **PR:** {pull_request_url}

## Label
**{helpful | unhelpful | wrong}**

## Reasoning
{reasoning from step 04}

5b. severity.md

# Severity: {id}

- **Comment:** {first 80 chars of body}...
- **File:** {file_path}:{diff_line}
- **PR:** {pull_request_url}

## Label
**{nit | moderate | critical}**

## Reasoning
{reasoning from step 05}

5c. context_scope.md

# Context Scope: {id}

- **Comment:** {first 80 chars of body}...
- **File:** {file_path}:{diff_line}
- **PR:** {pull_request_url}

## Label
**{diff | file | repo | external}**

## Context Evidence
| diff_line | file_path | why |
|---|---|---|
| {line} | {path} | {reason} |

## Reasoning
{reasoning from step 06}

5d. advanced.md

# Advanced: {id}

- **Comment:** {first 80 chars of body}...
- **File:** {file_path}:{diff_line}
- **PR:** {pull_request_url}

## Label
**{Repo-specific conventions | Context outside changed files | Recent language/library updates | Better implementation approach | False}**

## Reasoning
{reasoning from step 07}

7. Update task_info.md

Add to the Output section:

## Output
- **labels.json:** deliverables/labels.json
- **context.json:** deliverables/context.json
- **quality.md:** deliverables/quality.md
- **severity.md:** deliverables/severity.md
- **context_scope.md:** deliverables/context_scope.md
- **advanced.md:** deliverables/advanced.md
- **Generated:** {timestamp ISO 8601}
- **Validation:** PASSED / PASSED WITH WARNINGS

8. Update progress

Update progress.md: step 08 status = "done", Completed = {timestamp ISO 8601}.

Set Current Step to: "ALL COMPLETE"

9. Final message

Display:

== Task {id} COMPLETE ==
All 8 steps completed.

Output files:
  - deliverables/labels.json (for submission)
  - deliverables/context.json (context table for platform)
  - deliverables/quality.md
  - deliverables/severity.md
  - deliverables/context_scope.md
  - deliverables/advanced.md

Labels summary:
  Quality:       {value}
  Severity:      {value}
  Context Scope: {value}
  Advanced:      {value}