prompt-budget-preflight
Pre-flight check to run before any Agent-tool spawn. Prevents the "Prompt is too long" rejection class by enforcing a 5-section minimum-viable-prompt template and estimating forwarded context cost. Use whenever you are about to call the Agent tool with more than a one-paragraph prompt.
What this skill does
# Prompt Budget Preflight
Run this **before** calling the Agent tool. It takes under 30 seconds and prevents the most
common subagent failure mode: rejection with "Prompt is too long".
## Why this exists
Subagents inherit a large base context before your prompt is even appended:
- Global `~/.claude/CLAUDE.md`
- Project `.claude/CLAUDE.md` chain (including any rules files referenced via `@`)
- Any auto-loaded plugin `CLAUDE.md` in the subagent's `cwd`
- `MEMORY.md` and auto-memory entries
- System prompts for the selected `subagent_type`
- Tool definitions for tools the subagent can see
In our own telemetry from one 9-spawn session:
| Spawn type | Prompt size | Result |
|---|---|---|
| `Explore` × 2 (generic) | ~900 words each | Both **rejected** — "Prompt is too long", 0 tokens used, 2.3-2.7s round-trip |
| Named specialists × 7 (architecture / perf / security / dx / ux / researcher / code-reviewer) | ~200-400 words each | **All 7 succeeded** |
Reject rate: 22%. Every rejection is a wasted ~3 seconds and a wasted decision cycle.
## The 5-section minimum-viable prompt
Target **~300 words total**. If you exceed 800 words you almost certainly have dead weight.
```
1. TARGET
- Absolute file paths or directory scope
- Clear in/out-of-scope boundaries
2. TASK
- One sentence: what to produce
- Read-only vs write explicitly stated
3. FORMAT
- Output shape (YAML, table, bullets)
- Word ceiling or row cap
- Any required fields / schema
4. CONSTRAINTS
- What NOT to do
- Paths / patterns / subdirectories to avoid
- Word-limit enforcement hint
5. RULES (optional, only when applicable)
- Plugin-specific guardrails
- "Do not follow interview-first workflow — this is an audit"
- "Do not write files — this is analysis only"
```
## Preflight checklist
Before each `Agent` call, answer:
1. **Subagent type**: am I using a named specialist that already has the relevant system prompt, or a generic (`Explore`, `general-purpose`)? **Prefer specialists** — see `skills/agentic-patterns/SKILL.md` Part 4.
2. **Prompt length**: is my prompt under 500 words? If over 800, trim.
3. **Enumeration test**: am I re-listing facts the agent can grep? If yes, drop them — say "grep for X" instead of pasting X.
4. **Context inheritance**: will the agent spawn inside a directory whose `CLAUDE.md` is large (>4KB) and unrelated to the task? If yes, include a `RULES` line telling it to ignore that file's workflow prescriptions.
5. **Output ceiling**: did I specify a word ceiling and/or row cap? If no, add one.
## Failure-mode playbook
| If agent rejects with… | Do this |
|---|---|
| "Prompt is too long" | Cut enumerated facts; swap generic subagent for a named specialist; add RULES line telling it to skip directory CLAUDE.md workflows |
| Empty or stub output | Prompt was too vague — add concrete file:line references and one worked example of the output shape |
| Off-topic ramble | Missing CONSTRAINTS section — add explicit "do NOT explore X / Y / Z" |
| Timeout (>5min on simple task) | Missing word ceiling + missing scope boundary — add both |
## Template — copy-paste to start
```
You are analyzing <SCOPE>.
TARGET: <absolute paths>
TASK: <one sentence; read-only or write>
FORMAT:
<schema / shape>
Keep under <N> words.
CONSTRAINTS:
- Do not <X>
- Do not explore <Y>
RULES (if applicable):
- Ignore the <plugin-name> CLAUDE.md's interview-first workflow;
this is an audit, not a setup run.
```
## When to skip the preflight
- Prompt is under 2 paragraphs AND the subagent type matches a named specialist.
- The task is so scoped a wrong output won't cost a retry (e.g., "find all files matching X").
Everything else: run the 5-question checklist above. It's cheaper than a rejected spawn.
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.