agentleFS
qyndexqyndex/claude-init/.claude/skills/context-budget/SKILL.md

context-budget

Skill2 starsChanged 5 months agoMITView on GitHub

Inventory and recommend prunes for loaded agents, skills, MCP servers, and rules. Reports always-needed / sometimes-needed / rarely-needed buckets with token costs. From affaan-m/ECC context-budget.

57 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use agent-payment-x402, openclaw-persona-forge and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • TironFoxSkill

    agent-payment-x402

    Add x402 payment execution to AI agents — per-task budgets, spending controls, and non-custodial wallets via MCP tools. Use when agents need to pay for APIs, services, or other agents.

    Also in 6 projects that use this

  • TironFoxSkill

    openclaw-persona-forge

    为 OpenClaw AI Agent 锻造完整的龙虾灵魂方案。根据用户偏好或随机抽卡, 输出身份定位、灵魂描述(SOUL.md)、角色化底线规则、名字和头像生图提示词。 如当前环境提供已审核的生图 skill,可自动生成统一风格头像图片。 当用户需要创建、设计或定制 OpenClaw 龙虾灵魂时使用。 不适用于:微调已有 SOUL.md、非 OpenClaw 平台的角色设计、纯工具型无性格 Agent。 触发词:龙虾灵魂、虾魂、OpenClaw 灵魂、养虾灵魂、龙虾角色、龙虾定位、 龙虾剧本杀角色、龙虾游戏角色、龙虾 NPC、龙虾性格、龙虾背景故事、 lobster soul、lobster character、抽卡、随机龙虾、龙虾 SOUL、gacha。

    Also in 5 projects that use this

  • loulanyueSkill

    nuxt4-patterns

    Nuxt 4 app patterns for hydration safety, performance, route rules, lazy loading, and SSR-safe data fetching with useFetch and useAsyncData.

    Also in 9 projects that use this

  • renepfSkill

    orch-pipeline

    Shared orchestration engine for the orch-* skill family. Defines the gated Research-Plan-TDD-Review-Commit pipeline, the size classifier, the agent map, and the two human gates that the orch-* operation skills delegate to. Not usually invoked directly.

    Also in 4 projects that use this

  • Exploration-labsSkill

    token-budget-advisor

    Proactive token budget assessment and task chunking strategy. Use this skill when queries involve multiple large file uploads, requests for comprehensive multi-document analysis, complex multi-step workflows with heavy research (10+ tool calls), phrases like "complete analysis", "full audit", "thorough review", "deep dive", or tasks combining extensive research with large output artifacts. This skill helps assess token consumption risk early and recommend chunking strategies before beginning work.

    Also in 11 projects that use this

  • loulanyueSkill

    santa-method

    Multi-agent adversarial verification with convergence loop. Two independent review agents must both pass before output ships.

    Also in 8 projects that use this

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 53 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Don't archive without invocation log: Use .claude/hooks/.log/skill.log (when skill-router records triggers) as ground truth.

    line 87
  2. Distinctive:

    Never auto-archive agents: Always confirm with the user.

    line 88

“Common” is measured: kept by at least 6 in 10 of the 53 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Context Budget

Most agents accumulate skills/agents/rules cruft over time. This skill audits the current load and recommends prunes.

What it scans

SourcePathCost source
Agents.claude/agents/**/*.mdFrontmatter loaded at startup; body loaded on dispatch
Skills.claude/skills/*/SKILL.mdFrontmatter loaded; body on trigger
Rules.claude/rules/*.md (if any)Loaded by path-glob match
MCP servers.mcp.jsonTool schemas; deferred unless alwaysLoad: true
Memory.claude/memory/MEMORY.md + topic filesFirst 200 lines auto-loaded
Output styles.claude/output-styles/*.mdActive one loaded into system prompt

What it computes

For each item:

  • Token count (approximate via wc -c / 4)
  • Loaded? (yes/no based on alwaysLoad / matched paths / active status)
  • Last invoked (from .claude/hooks/.log/)
  • Recommendation: keep / defer / archive

Report format

# Context Budget Report — <date>

## Always-loaded (cost: <N> tokens)
- .claude/CLAUDE.md                       ████ 1840 tok    keep
- .mcp.json: filesystem (alwaysLoad)      ██ 920 tok        keep
- .mcp.json: git (alwaysLoad)             ██ 880 tok        keep
- .mcp.json: github (alwaysLoad)          ███ 1320 tok      keep

## Sometimes-loaded (cost on trigger: <N> tokens average)
- skills/autopilot                        body 4200 tok    invoked 12x this week    keep
- skills/self-heal                        body 1800 tok    invoked 4x                keep
- skills/research                         body 1200 tok    invoked 1x                consider deferring (low usage)

## Never invoked (cost: 0 active, <N> at startup for frontmatter)
- skills/onboard                          frontmatter only   not invoked in 30d        consider archiving
- agents/specialists/doc-writer           frontmatter only   not invoked in 14d        keep (low cost)

## MCP tool schemas (with Tool Search)
- chrome-devtools (deferred)               0 tok at start, ~3000 on first use
- playwright (deferred)                    0 tok at start
- apify (deferred)                         0 tok at start, ~8000 on first search
...

## Total
Startup cost: ~12,400 tokens (CLAUDE.md + agents + skill frontmatter + alwaysLoad MCPs)
Typical session: ~28,000 tokens (after a few skills triggered)
Budget headroom on 1M context: ~972,000 tokens
Recommendation: ✓ healthy

Recommendations

  • Defer (alwaysLoad: false on MCPs, or skill stays in dir but description trimmed)
  • Archive (move to .claude/skills/.archive/ — not loaded by Claude Code)
  • Trim (rewrite a verbose skill body to <500 lines)
  • Consolidate (two similar skills → one)

Manual usage

/context-budget          # generate the report to stdout
/context-budget --archive-stale     # auto-archive items not used in 60d
/context-budget --by-cost           # sort by token cost descending

Cache facts (gap-audit G1/G5)

  • "Cache hit rate dropping" is now measurable: bash .claude/scripts/cost-report.sh day writes cache_hit_rate into .claude/hooks/.log/cost-summary.json; the statusline shows cache:N% and harness-doctor warns below CACHE_HIT_MIN (default 40%).
  • TTL is ~5 minutes and not configurable from Claude Code. Turns spaced >5min apart re-pay the full cache write — keep unattended loops tight, and expect the first turn after an idle break to be cache-creation, not cache-read.

Hard rules

  • Don't archive without invocation log. Use .claude/hooks/.log/skill.log (when skill-router records triggers) as ground truth.
  • Never auto-archive agents. Always confirm with the user.
  • Stale ≠ useless. Some agents (release, security) are rarely-but-critically used.

References

  • affaan-m/ECC/skills/context-budget (source)
  • Anthropic's progressive disclosure pattern (frontmatter cheap, body lazy)
  • Claude Code Tool Search (deferred schemas)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.