loop-author-a
Write high-quality session prompts (SESSION_WORKER.md and SESSION_REVIEWER.md) for an agent loop. Reads repo context and follows the loop-author guide to produce targeted, effective prompts. Use when setting up a new loop or rewriting session prompts for an existing one.
What this skill does
> **Action skill** — Write session prompts for agent loops following the loop-author guide.
## Input
$ARGUMENTS
---
Write SESSION_WORKER.md and SESSION_REVIEWER.md for an agent loop. These prompts define what the loop accomplishes and how the reviewer evaluates completion.
## Gather Context
Before writing, read these sources to understand the repo and task:
### 1. Repo Context
```bash
# Project instructions
cat CLAUDE.md 2>/dev/null
cat .claude/CLAUDE.md 2>/dev/null
# Recent work
git log --oneline -20
# Repo structure (top-level)
ls -la
```
### 2. Task Context
From `$ARGUMENTS`, extract:
- **What** the loop should accomplish
- **Where** in the codebase the work lives
- **Why** this work matters (if stated)
- **Constraints** (language, patterns, dependencies)
### 3. Loop Author Guide
Follow the session prompt anatomy from `agents/prompts/loop-author.md` (or `~/claude-tooling/agents-src/prompts/loop-author.md`). The four required sections are:
1. **Objective** — single paragraph, no ambiguity
2. **Context** — key files, existing patterns, constraints
3. **Behavior Targets** — observable goals, not implementation steps
4. **Acceptance Criteria** — verifiable conditions for STOP
## Write SESSION_WORKER.md
```markdown
# Objective
{One paragraph. What is the end goal? What is the output?}
{Read X, Y, Z before starting — point to specific files.}
## Context
{What exists today. Key files and directories. Existing patterns to follow.
What has been tried. Hard constraints.}
## Behavior Targets
When this is done, the following should be true:
1. {Observable behavior — written from a tester's perspective}
2. {Observable behavior}
3. ...
## Acceptance Criteria
- {Verifiable condition — the minimum bar}
- {Verifiable condition}
- ...
## Constraints
- {Hard constraint — language, patterns, dependencies}
- {Hard constraint}
```
### Worker Prompt Principles
- **Objective is the north star.** If the objective is vague, the loop wanders. Be specific.
- **Point to files, not concepts.** "Read `src/auth/middleware.ts`" not "understand the auth system."
- **Behavior targets are observable.** "Users can log in with SSO" not "Add OAuthProvider class."
- **No decomposition.** Don't tell the worker how to break down the work — that's what TODO.md is for.
- **No code snippets.** Session prompts are strategic, not tactical.
## Write SESSION_REVIEWER.md
```markdown
# Review Brief
You are reviewing progress toward: {one-line objective summary}
## What to Verify
- {Verification point — what to check at each iteration}
- {Verification point}
## Quality Bar
- {What good looks like for this specific task}
- {Standards that apply — tests, linting, patterns}
## Watch For
- {Known risk or failure mode}
- {Common mistake for this type of work}
## Done Condition
Write STOP.txt when:
- {Specific, observable condition}
- {Specific, observable condition}
```
### Reviewer Prompt Principles
- **Same objective, different perspective.** The reviewer sees the same goal but evaluates rather than builds.
- **Shorter than the worker prompt.** Reviewers need clarity, not volume.
- **Concrete done conditions.** "All tests pass" not "code is clean."
- **Watch-for items are specific.** "Over-engineering the error handling" not "bad code."
## Output Location
Write the prompts to:
```
agents/loops/{loop-id}/session/SESSION_WORKER.md
agents/loops/{loop-id}/session/SESSION_REVIEWER.md
```
Create the directory structure if it doesn't exist:
```bash
mkdir -p agents/loops/{loop-id}/session
```
## After Writing
1. Show the user both prompts for review
2. Ask if they want to adjust anything before launch
3. If they're satisfied, suggest launching with `loop-launch-a` or directly:
```bash
./agents/run.sh --loop-id {loop-id}
```
## Common Mistakes to Avoid
| Mistake | Fix |
|---------|-----|
| Vague objective | One paragraph, no ambiguity — name the output |
| Steps instead of goals | Describe *what*, not *how* |
| No acceptance criteria | List verifiable conditions |
| Reviewer copies worker | Write from reviewer perspective — it's a review brief |
| Too much context | Only include what's needed to start work |
| Implementation approach in prompt | Let the worker decide, reviewer course-correct |
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.