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.
npx agentmods add skills/hehee9/codex-compressor/installnpx skills add hehee9/codex-compressor --skill installgit clone --depth 1 https://github.com/hehee9/codex-compressorWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00014 | $0.00728 |
| Opus 5 | $0.00007 | $0.00364 |
| Sonnet 5 | $0.00003 | $0.00146 |
| Haiku 4.5 | $0.00001 | $0.00073 |
Grade A, and why
install 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 2d ago.
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.
What it actually says
설치
플러그인은 소스 저장소가 삭제된 뒤에도 동작해야 하므로, 모든 명령은 설치된 플러그인의
runtime을 사용합니다. ${PLUGIN_ROOT}는 Codex가 이 플러그인에 제공하는 설치 경로입니다.
codex-compressor는 두 가지 중 한 가지 모드만 선택해 사용합니다.
- 설정 관리 모드: 플러그인 없이 저장소 루트의
install.ps1또는install.sh를 실행합니다. - 플러그인 후크 모드: 아래의 로컬 마켓플레이스에서 플러그인을 설치한 뒤 설정 관리 설치 단계를 한 번 실행합니다.
두 모드를 동시에 활성화하면 설정을 서로 덮어쓸 수 있으므로 함께 사용하지 않습니다. 새 설치의 버퍼 기본값은 16384입니다. 기존 설정에서 도구가 인식하는 155200 버퍼를 발견하면 그 값을 유지하고, 인식하지 못한 값은 임의로 바꾸지 않고 사용자에게 확인을 요청합니다.
설정 관리 모드
# macOS/Linux
PYTHONPATH="${PLUGIN_ROOT}/runtime" python3 -m codex_compressor install --mode standalone
PYTHONPATH="${PLUGIN_ROOT}/runtime" python3 -m codex_compressor status
PYTHONPATH="${PLUGIN_ROOT}/runtime" python3 -m codex_compressor doctor
# Windows PowerShell
$env:PYTHONPATH = Join-Path $env:PLUGIN_ROOT "runtime"
py -3 -m codex_compressor install --mode standalone
py -3 -m codex_compressor status
py -3 -m codex_compressor doctor
소스 저장소에서 처음 설치하는 경우에는 저장소 루트의 설치 도우미를 사용합니다.
# macOS/Linux
./install.sh install --mode standalone
# Windows PowerShell
.\install.ps1 install --mode standalone
설치 전에 원본 설정을 백업하고, 원본 및 사용자가 편집한 모든 키를 보존하는지 확인합니다.
플러그인 후크 모드
소스 저장소에서 처음 설치할 때는 루트의 설치 도우미가 src 레이아웃을 준비합니다.
# macOS/Linux
./install.sh install --mode plugin
# Windows PowerShell
.\install.ps1 install --mode plugin
그 다음 저장소 루트에서 마켓플레이스를 등록합니다.
codex plugin marketplace add .
codex plugin marketplace list
codex plugin add codex-compressor@codex-compressor-local
플러그인 설치만으로 설정 관리 단계가 자동 실행되지는 않습니다. 설정 관리 단계가 필요하면 설치된 런타임으로 다음을 실행합니다.
# macOS/Linux
PYTHONPATH="${PLUGIN_ROOT}/runtime" python3 -m codex_compressor install --mode plugin
# Windows PowerShell
$env:PYTHONPATH = Join-Path $env:PLUGIN_ROOT "runtime"
py -3 -m codex_compressor install --mode plugin
Codex에서 /hooks를 열어 후크 신뢰 승인을 완료한 뒤 새 세션에서 status와 doctor를 실행합니다.
What ships with it
1 file 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.
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.
- 2d ago First seen · 77 lines · 14 tokens per session scan A c55112a92938
install is a skill published in the GitHub repository hehee9/codex-compressor (10 stars, last pushed 7d ago), licensed Apache-2.0. It adds 14 tokens to every session and 728 once invoked, about $0.0001 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 2026-08-31.
Other skills, from other repositories
writing-skills
Use this whenever the user asks to write, author, or scaffold a new SecureContext skill (Anthropic-style filesystem skill at /.claude/skills/ / or /.claude/skills/ /). The skill walks the agent through the four-invariant checklist, generates a SKILL.md from a template, and runs the bundled linter + admission preview…
publish-github-release
Use this whenever the user asks to ship, publish, release, tag, or cut a new version of the project — OR when CHANGELOG.md has an "Unreleased" / in-progress section ready to be shipped. The skill bumps the version across all version-bearing files, regenerates the CHANGELOG entry from git log since the last tag…
ctxwise
Audit and optimize OpenAI Codex context locally, compile model and subagent profiles, detect capability drift, create redacted lockfiles, inspect model-visible prompt structure, and render honest token, credit, quota, or API-cost receipts. Use when the user explicitly invokes CtxWise or asks to run its installed CLI…
pf-handoff
Чекпоинт или закрытие сессии — зафиксировать состояние, ничего не потеряв. Checkpoint or close a session — reconcile what is proven, rewrite the live HANDOFF state file (§13), statuses and journal. NOT for handing work to another agent or worktree (that is orca-cli/orchestration). Use when the context window is…
pf-resume
Подхватить работу прошлой сессии — продолжить как в том же чате. Picking up work from the live HANDOFF file (§13) — continue in a new chat or after /clear as if it were the same chat. Use when «/pf-resume», «продолжи с прошлой сессии», «подхвати работу», «восстанови состояние», resume previous session state.
minimize-token-usage
Reduce Codex context and token use while preserving evidence and correctness. Use for long coding, research, debugging, repository exploration, noisy logs, large command or MCP outputs, repeated file reads, cost-sensitive work, or when the user asks for concise, cheap, low-token, or context-efficient execution. Also…