Nm Imbue Structured Output

Format final review deliverables with consistent structure for comparable findings across reviews

Security Scan

This is an instruction-only formatting helper that is internally consistent with its stated purpose and requests only a single config path for evidence logs; nothing requested appears disproportionate or unexpected.

Purpose & Capability

The skill's name and description (formatting structured deliverables) match the SKILL.md steps: selecting templates, formatting findings, assigning actions, and attaching an appendix. The declared required config path (night-market.imbue:proof-of-work) is coherent with the appendix/evidence references.

Instruction Scope

All instructions are about formatting and assembling deliverables. The SKILL.md references the 'proof-of-work' evidence log and repeatedly suggests verifying commands with '--help' but does not name or run specific unexpected commands or request arbitrary files or environment variables. Some guidance is generic/vague (e.g., 'run the command with --help' without naming the command) but this is scope-consistent for a template/formatting skill.

Install Mechanism

No install spec and no code files — instruction-only. There is nothing downloaded or written to disk by the skill itself in the provided material.

Credentials

The skill requires no environment variables or credentials. The single required config path (night-market.imbue:proof-of-work) is consistent with the skill's purpose of including evidence logs in an appendix; it does not request unrelated secrets or broad system access.

Persistence & Privilege

always:false and default model invocation settings are used. The skill does not request persistent privileges or modify other skills' configurations in the provided instructions.

Assessment

This is a formatting/template helper and appears safe to install. Before enabling it, confirm what is stored under the configured path night-market.imbue:proof-of-work (it will be referenced for the Appendix) so you know whether sensitive data might be included. Also verify that any commands your agent might run (the doc suggests running '--help' for verification) are the expected tooling in your environment. If the skill later requests credentials, broader file paths, or an always:true setting, re-evaluate before installing.

Night Market Skill — ported from claude-night-market/imbue. For the full experience with agents, hooks, and commands, install the Claude Code plugin.

Table of Contents

• When to Use
• Activation Patterns
• Required TodoWrite Items
• Step 1: Select Template (structured-output:template-selected)
• Step 2: Format Findings (structured-output:findings-formatted)
• Step 3: Assign Actions (structured-output:actions-assigned)
• Step 4: Attach Appendix (structured-output:appendix-attached)
• Output Quality Checklist
• Exit Criteria

Structured Output

When To Use

• When finalizing any review or analysis.
• To format findings in a consistent and actionable way.
• Before presenting results to stakeholders or committing them to documentation.

When NOT To Use

• Capturing evidence during analysis - use proof-of-work
• Reviewing changes - use diff-analysis or review-core first

Activation Patterns

Trigger Keywords: format, structure, deliverable, report, organize, present, consistent Contextual Cues:

• "format this as a report" or "structure the output"
• "create a deliverable" or "present these findings"
• "organize this consistently" or "standardize the format"
• "make this actionable" or "prepare for stakeholders"

Auto-Load When: Finalizing any analysis deliverable or when consistent formatting is requested.

Required TodoWrite Items

1. structured-output:template-selected
2. structured-output:findings-formatted
3. structured-output:actions-assigned
4. structured-output:appendix-attached

Mark each item complete as you finish the corresponding step.

Step 1: Select Template (structured-output:template-selected)

• Choose output format based on deliverable type:
  • Review Report: Summary, Findings, Recommendations, Evidence.
  • PR Description: Summary, Changes, Test Plan, Notes.
  • Release Notes: Highlights, Breaking Changes, Fixes, Credits.
  • Incident Report: Timeline, Impact, Root Cause, Remediation.
• Confirm audience and required detail level.

Step 2: Format Findings (structured-output:findings-formatted)

• Use consistent finding structure:

  ### [SEVERITY] Finding Title
  **Location**: file.rs:123
  **Category**: Security | Performance | Correctness | Style
  **Description**: Brief explanation of the issue.
  **Evidence**: [E1, E2] - Reference to evidence log.
  **Recommendation**: Specific remediation steps.
  

  Verification: Run the command with --help flag to verify availability.
• Severity levels: CRITICAL, HIGH, MEDIUM, LOW, INFO.
• Order findings by severity, then by file location.

Step 3: Assign Actions (structured-output:actions-assigned)

• Convert findings to actionable items:

  ## Action Items
  - [ ] [HIGH] Fix SQL injection in auth.py:45 (@security-team, P1)
  - [ ] [MEDIUM] Add input validation to API endpoint (@backend, P2)
  - [ ] [LOW] Update deprecated dependency (@devops, P3)
  

  Verification: Run the command with --help flag to verify availability.
• Include owner assignment where known.
• Add priority indicators (P1/P2/P3) for triage.
• Note dependencies between actions.

Step 4: Attach Appendix (structured-output:appendix-attached)

• Compile supporting materials:

  ## Appendix
  ### A. Commands Run
  [Full evidence log from imbue:proof-of-work]
  
  ### B. External References
  [Citations and documentation links]
  
  ### C. Raw Data
  [Large outputs, full diffs, or data exports]
  

  Verification: Run the command with --help flag to verify availability.
• Keep main report concise; details in appendix.
• validate appendix is navigable with clear section headers.

Output Quality Checklist

Before finalizing:

• All findings have evidence references.
• Severity levels are justified.
• Recommendations are specific and actionable.
• No orphaned sections or placeholder text.
• Format renders correctly in target medium (GitHub, Confluence, etc.).

Exit Criteria

• Todos completed with formatted deliverable.
• Output follows selected template structure.
• Stakeholders can act on findings without clarification.

Troubleshooting

Common Issues

Command not found Ensure all dependencies are installed and in PATH

Permission errors Check file permissions and run with appropriate privileges

Unexpected behavior Enable verbose logging with --verbose flag