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 skills add cubxxw/agent-kit --skill manage-agent-kitgit clone --depth 1 https://github.com/cubxxw/agent-kitWrote 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.
[](https://agentmods.dev/skills/cubxxw/agent-kit/manage-agent-kit)<a href="https://agentmods.dev/skills/cubxxw/agent-kit/manage-agent-kit"><img src="https://agentmods.dev/badge/skills/cubxxw/agent-kit/manage-agent-kit/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.
<a href="https://agentmods.dev/skills/cubxxw/agent-kit/manage-agent-kit"><img src="https://agentmods.dev/badge/skills/cubxxw/agent-kit/manage-agent-kit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00093 | $0.01226 |
| Opus 5.5 | $0.00037 | $0.00490 |
| Sonnet 5.5 | $0.00019 | $0.00245 |
| Haiku 4.5 | $0.00009 | $0.00123 |
Grade A, and why
manage-agent-kit 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 130 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Manage Agent Kit
Treat the checked-out agent-kit repository as the reviewed source. Native
installs receive symlink views; an existing manager such as CC Switch can own
distribution instead. Choose one installation owner for each skill.
Start with evidence
- Identify the user's task, current host, and existing skill owner from runtime evidence and link destinations. Do not read private runtime state.
- Inspect
catalog.jsonand choose the smallest useful profile. For unspecified coding work usedeveloper, on the current host only. - Read references/operations.md for the selected installation path. If CC Switch owns the skills, preserve that owner.
- For a native checkout run
doctor --strict, thenplan --profile <profile> --tool <host> --json. Explain the result the user will get and every conflict before applying the authorized setup.
Do not silently replace a real directory or a symlink owned by another tool. Report the exact conflict and preserve it.
Install or repair links
Use the smallest applicable profile:
base: only this management skill.developer: source-grounded engineering and agent infrastructure.design: recursive direction branching, official frontend guidance, UI/UX evidence, and frontend preflight.diagramming: collaborative Excalidraw exploration and validated Archify technical documentation.prototyping: one uncertain decision, tested with observable cases.writing: optional first-person Threads drafting and confirmed publishing.full-stack: explicitly combine development, design, diagramming, and prototyping.top: every accepted skill, including optional writing; choose explicitly.all: compatibility alias fortop.
Preview first:
./bin/agent-kit plan --profile <profile> --tool <host> --json
The plan is read-only and includes ready, actions, and conflicts.
With conflicts, return the complete explanation and the next action. Known
conflicts block the whole install before any link is created. Then install
the selected profile and verify host discovery and one small task using
the repository's first-run guide. A same-named file or symlink is not task
success. Use --replace-managed only for stale symlinks that already
point into this checkout. It must never replace copied files or foreign links.
What ships with it
3 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.
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.
- 4d ago Changed · +3 lines 81d7d489396e
- 22d ago Changed · +14 lines ba02c575881c
- 1mo ago First seen · 113 lines · 93 tokens per session scan A 802c66eec5d8
manage-agent-kit is a skill published in the GitHub repository cubxxw/agent-kit (2 stars, last pushed 4d ago), licensed MIT. It adds 93 tokens to every session and 1,226 once invoked, about $0.0004 per session on Opus 5.5. 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
fieldops-ctf-reverse
Reverse engineering of compiled, packed, virtualized, or obfuscated targets for authorized CTF challenges. Use when the task is to understand, decompile, disassemble, or unpack an unknown executable across ELF, PE, Mach-O, APK, WASM, firmware, bytecode, or custom-VM targets.
CCSwitch-operations
A set of operating instructions for safely changing and synchronising CC Switch, an app that manages prompts, skills, MCP servers, providers, and Codex model settings across several AI coding applications.
fieldops-ctf-writeup
Generate a submission-style writeup for a solved CTF challenge. Use after a solve, when the user wants a reproducible competition report covering metadata, ordered solution steps, the solve script, and the recovered flag.
fieldops-powershell-utf8
Writes, reviews, and repairs PowerShell so Persian/Farsi and other Unicode text survives console input, console output, native-process pipelines, and file I/O. Use when text turns into mojibake or question-mark boxes, when a UTF-8 BOM must be added or removed, for multilingual or non-ASCII automation, or when a script…
fieldops-ctf-operator
Triage, route, and validate authorized CTF investigations. Use when the category is unknown, the user is stuck, evidence conflicts, or a claimed solve must be reproduced from a clean baseline. Dispatches to the most relevant fieldops-ctf- specialist.
fieldops-prompt-decorators
Interprets and executes Prompt Decorators written with +++Name(key=value) syntax, handling message vs. chat scope, parameter validation, decorator composition, conflict resolution, meta-decorator expansion (N2 and Storm), retained-state inspection (ActiveDecs and AvailableDecs), selective clearing, and conversation…