adaptive-compaction

adaptive-compaction is a skill for Claude Code from Yrzhe/claude-skills. It costs 217 tokens per session (3,024 once invoked), scanned A, original, MIT.

A set of rules for deciding when a coding agent should shorten its conversation history, save a snapshot, start a separate conversation, or ask first. Conversation compaction means replacing older messages with a shorter summary to make room for new work.

In plain words
What is it for?
It is for managing context limits across Claude Code, Codex, and MCP-compatible hosts, including pruning, snapshots, compaction advice, and session forks.
Why use it?
It helps avoid losing important context or compacting at a bad time, especially when the conversation is long or the topic changes.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: mentions Claude Code; mentions Codex.

Part of the adaptive-compaction plugin — 1 skill shipped together

Good fit It is for managing context limits across Claude Code, Codex, and MCP-compatible hosts, including pruning, snapshots, compaction advice, and session forks.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/yrzhe/claude-skills/adaptive-compaction
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add Yrzhe/claude-skills --skill adaptive-compaction
Clone the repo
git clone --depth 1 https://github.com/Yrzhe/claude-skills

Made for: Claude Code.

Or install adaptive-compaction, the plugin that ships this one along with the rest of its 1 skill.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for adaptive-compaction

README.md
[![agentmods](https://agentmods.dev/badge/skills/yrzhe/claude-skills/adaptive-compaction/github.svg)](https://agentmods.dev/skills/yrzhe/claude-skills/adaptive-compaction)
Your own site
<a href="https://agentmods.dev/skills/yrzhe/claude-skills/adaptive-compaction"><img src="https://agentmods.dev/badge/skills/yrzhe/claude-skills/adaptive-compaction/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for adaptive-compaction

Your own site · 80×15
<a href="https://agentmods.dev/skills/yrzhe/claude-skills/adaptive-compaction"><img src="https://agentmods.dev/badge/skills/yrzhe/claude-skills/adaptive-compaction.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 217 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,024 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00217 $0.03024
Opus 5 $0.00109 $0.01512
Sonnet 5 $0.00043 $0.00605
Haiku 4.5 $0.00022 $0.00302

Measured 12d ago against content hash 5bf20ef886a5, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

adaptive-compaction scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.

The scan reads SKILL.md. This mod also ships 10 executable files (assets/hooks/claude-code/pre-compact.sh, assets/hooks/claude-code/session-start.sh, assets/hooks/claude-code/user-prompt-submit.sh, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/adaptive-compaction/SKILL.md · 263 lines

How it starts

The opening of the file, as written. The whole thing — 263 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Adaptive Compaction

Use this skill to install, inspect, or operate the adaptive compaction policy and recovery layer for coding agents. Keep the layer add-on only: decide when to preserve, prune, compact, fork, or ask; never rewrite the host agent core.

Operating Rules

Follow these rules before taking action:

  1. Preserve state before any risky compaction path.
  2. Run zero-LLM pruning before any LLM summary recommendation.
  3. Treat relevance as advisory. Never use low relevance alone for destructive action.
  4. Prefer fork/reset on topic shift when headroom is healthy.
  5. Enforce FORCED_COMPACT only on hook-rich hosts. Degrade Codex and MCP-only hosts to advise/persist.
  6. Keep stable instructions before volatile state to avoid prompt-cache churn.
  7. Log every decision with thresholds and host capability tier.
  8. Use portable paths such as <skill-dir>/scripts/decide.py; do not hardcode install paths.
  9. When the per-turn additionalContext reports a non-NOOP action, lead your next reply by surfacing it to the user in their language (forced compaction needed, fork recommended, decision needs confirmation). Then continue with the user's actual request. Do not silently swallow these signals — that is the only channel by which the skill can ask the user to act.

Load References Conditionally

Do not read every reference by default. Load only the rows needed for the current request.

Task Read
Check JSON contracts or script I/O <skill-dir>/references/contract.md
Explain or tune the four-signal decision table <skill-dir>/references/decision-policy.md
Make or review host portability claims <skill-dir>/references/capability-matrix.md
Validate release safety or risk coverage <skill-dir>/references/risk-acceptance-tests.md
Plan trace evaluation or v1 to v2 gates <skill-dir>/references/eval-protocol.md
Install Claude Code hooks <skill-dir>/assets/hooks/claude-code/settings.json and the three shell scripts in that folder
Install Codex hooks/config <skill-dir>/assets/hooks/codex/hooks.json and <skill-dir>/assets/hooks/codex/config.toml
Run deterministic fixture checks <skill-dir>/scripts/fixtures/*.json

Read the full file on GitHub · 263 lines

Files

What ships with it

30 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 12d ago First seen · 263 lines · 217 tokens per session scan A 5bf20ef886a5

Subscribe to this mod's changes

adaptive-compaction is a skill published in the GitHub repository Yrzhe/claude-skills (33 stars, last pushed 3mo ago), licensed MIT. It adds 217 tokens to every session and 3,024 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

media-ingest

Ingest video, audio, PDF, book, screenshot, and GitHub repo content into the brain. Multi-format handling with entity extraction and backlink propagation. Covers video-ingest, youtube-ingest, and book-ingest subtypes.

garrytan/gbrain · 52 tokens

mem0-oss-to-platform

Plan and then execute a migration of a project from the mem0 open-source / self-hosted SDK (the local Memory class) to the mem0 Platform / hosted / managed SDK (the MemoryClient class). Use this whenever a developer wants to move, switch, or migrate their mem0 usage off OSS/self-hosted to the hosted API — e.g.…

mem0ai/mem0 · 273 tokens

Cortex

Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…

danielmiessler/LifeOS · 196 tokens

memory

Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.

opensquilla/opensquilla · 44 tokens

ha-data-stores

Map of Hope Agent's local data stores and safe read-only query workflow. Use when the user asks where Hope Agent stores data, wants to inspect sessions/messages/memory/logs/background jobs/knowledge indexes/settings, asks the model to query local app data, or debugging requires checking persisted state. Trigger…

shiwenwen/hope-agent · 115 tokens

establishing-project-context

Use when the user asks to establish shared project language, or project work exposes a conflicting, renamed, or deprecated domain term that needs active semantic modeling. Routine small tasks stay on the fast path.

GanyuanRan/Aegis · 45 tokens