backlog
Use when the user asks to "review backlog", "analyze backlog quality", "backlog health", or invokes /compass-rose:backlog. Reads all open and wontfix issue files and analyzes quality via the backlog-analyzer agent.
What this skill does
# Backlog Review Mode
You are now in **Backlog Review Mode**. Your role is to read all active issue files, assess their quality and readiness, and recommend the top 2-3 items to work on next based on priority, size, and definition quality.
## Your Focus
- **Issue discovery**: Glob all `.lore/work/issues/*.html` files
- **Status filtering**: Include items with status `open` or `wontfix` (exclude `done` and `closed`)
- **Quality analysis**: Spawn backlog-analyzer agent to assess definition quality
- **Recommendation**: Present 2-3 best options with detailed rationale
## Workflow
### 1. Discover Issue Files
Use Glob to find all issue files:
```
.lore/work/issues/*.html
```
If no files exist, stop and say:
```
No issue files found in .lore/work/issues/.
To get started, create your first issue with /compass-rose:add-item.
```
### 2. Parse Each Issue File
For each file, use Read to extract the meta tags from the `<head>` and the body text:
**Meta tags to extract**:
- `priority` — e.g., `P0`, `P1`, `P2`, `P3`
- `size` — e.g., `S`, `M`, `L`, `XL`
- `status` — e.g., `open`, `wontfix`, `done`, `closed`
- `date` — creation date
- `title` — issue title (also available as `<title>` tag)
**Body text**: Extract the readable text content from the HTML body.
### 3. Filter by Status
Include items where status is `open` or `wontfix`. Exclude items where status is `done` or `closed`.
`wontfix` items are intentionally included. They are deferred, not done, and should surface if context changes.
**If no items remain after filtering**:
```
No open or deferred issues found in .lore/work/issues/.
All issues are either done or closed.
To add new work, use /compass-rose:add-item.
```
### 4. Prepare Data for Agent Analysis
Build a JSON array from the filtered items:
```json
[
{
"filepath": ".lore/work/issues/fix-login-timeout.html",
"title": "Fix login timeout",
"priority": "P1",
"size": "S",
"status": "open",
"date": "2026-05-01",
"body": "<p>Users experiencing timeouts when submitting the login form...</p>"
}
]
```
### 5. Spawn Backlog Analyzer Agent
Invoke the backlog-analyzer agent with the prepared item data:
```
Analyze these project items and recommend the top 2-3 to work on next:
[Paste the formatted JSON array]
Focus on:
- Definition quality (clarity, completeness, acceptance criteria)
- Priority distribution
- Size balance (prefer a mix of quick wins and substantial work)
- Overall backlog health
Return your analysis with detailed rationale for each recommendation.
```
**Agent Reference**: `compass-rose/agents/backlog-analyzer.md`
The agent will:
1. Score each item on definition quality (0-10 scale)
2. Identify well-defined items (score 8-10)
3. Calculate recommendation scores (priority + size + quality)
4. Return top 2-3 recommendations with rationale
5. Provide backlog health summary
6. List items needing clarification
### 6. Present Agent Recommendations
Display the agent's analysis output directly to the user. The agent returns structured markdown with:
```markdown
# Backlog Analysis Results
**Items Analyzed**: [N total]
**Well-Defined Items**: [X items with score 8-10]
**Items Needing Clarification**: [Y items with score <5]
## Top Recommendations
### Recommendation 1: [Title] ([filename])
**Priority**: [P0/P1/P2/P3] | **Size**: [S/M/L/XL] | **Definition Quality**: [Well-Defined/Defined/Vague/Poorly Defined] ([score]/10)
**Rationale**:
- [Why this is recommended - link priority, size, definition quality]
- [What makes it ready to work on]
- [Any specific strengths]
**Definition Assessment**:
- **Clarity** ([0-3]/3): [Brief assessment]
- **Completeness** ([0-3]/3): [Brief assessment]
- **Acceptance Criteria** ([0-4]/4): [Brief assessment]
**File**: [filepath]
---
### Recommendation 2: [Title] ([filename])
[Same structure as Recommendation 1]
---
### Recommendation 3: [Title] ([filename]) [OPTIONAL]
[Same structure as Recommendation 1]
---
## Backlog Health Summary
**Priority Distribution**: [X P0, Y P1, Z P2, W P3]
**Size Distribution**: [A S, B M, C L, D XL]
**Definition Quality**:
- Well-Defined (8-10): [X items]
- Defined (5-7): [Y items]
- Vague (2-4): [Z items]
- Poorly Defined (0-1): [W items]
**Observations**:
- [Notable patterns]
- [Quality trends]
- [Recommendations for backlog improvement]
## Items Needing Clarification
1. **[Title]** ([filename]) - Score: [X]/10
- Missing: [What needs to be added]
- Suggest: [How to improve definition]
```
### 7. Handle Edge Cases
**All Items Poorly Defined**:
If the backlog-analyzer agent reports all items have low quality scores (<5), still present the top 2-3 but emphasize the need for clarification:
```
Warning: Most backlog items lack sufficient detail.
The recommendations below are based on priority and size only. Each item needs
clarification before implementation can begin. Consider adding:
- Clear problem or feature descriptions
- Reproduction steps (for bugs) or use cases (for features)
- Explicit acceptance criteria
Recommended next step: Clarify the highest-priority items before starting work.
```
**Missing Priority or Size Fields**:
```
Warning: Some items are missing priority or size fields.
Recommendations for those items will fall back to definition quality and date.
Consider setting priority and size in each issue file's meta tags.
```
## Requirements Mapping
This skill implements the following specification requirements:
- **REQ-CR-5**: Read and analyze file-based issue tracking from `.lore/work/issues/`
- **REQ-F-11**: Analyze backlog items and recommend based on priority, size, and definition quality
- **REQ-F-12**: Identify items that are well-defined (clear description and acceptance criteria)
- **REQ-F-13**: Present 2-3 options when asked for recommendations, with rationale
- **REQ-NF-3**: Handle missing custom fields gracefully (warn, don't fail)
- **REQ-NF-4**: Explain reasoning when making recommendations
## Anti-Patterns to Avoid
- **Don't query GitHub**: This skill reads local files only
- **Don't exclude wontfix**: These items are deferred, not done; the agent needs them
- **Don't fail silently**: If meta tags are missing, warn and continue with defaults
- **Don't bypass the agent**: Always use backlog-analyzer for quality assessment
- **Don't truncate agent output**: Present the agent's full analysis to maintain transparency
## Related Skills
- `/compass-rose:next-item` - Get an immediate recommendation for next work item (faster, simpler)
- `/compass-rose:add-item` - Create a new issue file
- `/compass-rose:reprioritize` - Codebase-aware priority updates
## Comparison to /compass-rose:next-item
| Aspect | /compass-rose:next-item | /compass-rose:backlog |
|--------|-------------------------|----------------------|
| **Speed** | Fast | Slower |
| **Scope** | Ready items only | All open and wontfix items |
| **Analysis** | Simple priority sort | Deep quality assessment |
| **Output** | Quick recommendation | Comprehensive backlog health report |
| **Use Case** | "What's next?" | "What's the state of my backlog?" |
## References
- **Spec**: REQ-CR-5, REQ-F-11, REQ-F-12, REQ-F-13, REQ-NF-3, REQ-NF-4
- **Agent**: `compass-rose/agents/backlog-analyzer.md` (quality scoring and recommendation)
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.