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 magic3007/dotfiles --skill toil-offloading-pigit clone --depth 1 https://github.com/magic3007/dotfilesWrote 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/magic3007/dotfiles/toil-offloading-pi)<a href="https://agentmods.dev/skills/magic3007/dotfiles/toil-offloading-pi"><img src="https://agentmods.dev/badge/skills/magic3007/dotfiles/toil-offloading-pi/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/magic3007/dotfiles/toil-offloading-pi"><img src="https://agentmods.dev/badge/skills/magic3007/dotfiles/toil-offloading-pi.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 | $0.00111 | $0.03271 |
| Opus 5.5 | $0.00044 | $0.01308 |
| Sonnet 5.5 | $0.00022 | $0.00654 |
| Haiku 4.5 | $0.00011 | $0.00327 |
Grade A, and why
toil-offloading-pi 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 28d 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 — 261 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Toil Offloading (pi)
The root agent is the strong, slow model. Subagents are cheap and fast. Design the work so the expensive model spends its budget on judgment, and the cheap models absorb volume.
The test is not "is this mechanical?" — it is "can this be cut off as a self-contained subtask with an observable acceptance condition?" If yes, it can be delegated, including implementation, refactoring, and writing tests. What stays with the root is the part that cannot be written down as a contract because it is the decision: which design is right, what the real requirement is, and whether the delivered work is acceptable.
Use the subagent tool exposed by the afang-subagent extension. Do not launch
agent CLIs (claude, codex, gemini, ...) as shell subprocesses.
The two-tier model
| Tier | Model | Owns |
|---|---|---|
| Root | the session's strong/slow model | Goal clarification, decomposition, architecture and design calls, contract authorship, integration, risk grading, acceptance |
| Subagents | cheap/fast (set by PI_SUBAGENT_PROVIDER + PI_SUBAGENT_MODEL) |
Executing contracts: recon, implementation, refactoring, data extraction, artifact generation, test writing, verification runs, review |
Delegation is the default for decomposable work because the marginal subtask costs a cheap model's time, not the root's. The root should be spending its context on the plan and the acceptance decision, not on typing out the changes.
Where the tier is declared
The split is one session-wide setting, not a per-agent choice:
export PI_SUBAGENT_PROVIDER="dspro"
export PI_SUBAGENT_MODEL="deepseek-flash[1m]"
It lives in ~/.common_shell_setup_local.sh so every pi session inherits it,
including the plain pi aliases. Resolution order in the afang-subagent
extension: these env vars → agent frontmatter model: → current session model.
Because the env vars win, do not pin models in agent frontmatter — that would be
a second source of truth that silently loses. To change the subagent tier, change
the env vars.
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.
- 28d ago First seen · 261 lines · 111 tokens per session scan A 7e640556489e
toil-offloading-pi is a skill published in the GitHub repository magic3007/dotfiles (11 stars, last pushed yesterday), licensed MIT. It adds 111 tokens to every session and 3,271 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 2026-09-18.
Other skills, from other repositories
dotfiles
Chezmoi dotfiles conventions and patterns. Use when modifying dotfiles, adding new managed files, working with chezmoi templates, or syncing agent skills and subagents through agentspec. Do NOT use for installing tools, running doctor, or updating a machine through the dotfiles CLI; use manage-machine.
use-notion
Reference for using the Notion API via the notion-agent subagent. Use to expore and update Notion workspace content.
orchestrate-agents
Orchestrate multiple agent CLIs (Claude, Codex, Antigravity) via tmux with a shared fleet store, dispatching one guardian subagent per pane. Survey-first: inspects and adopts existing tmux sessions, windows, and agent panes before creating anything new. Use when running a multi-agent session, dispatching parallel…
assess-quality
Foundational quality framework: the five questions (readable, easy to start, expands without bloat, consistent, intentional) every other dev skill is judged against, plus the dual-audience and workshop principles. Use when onboarding to a project, defining a quality bar, setting an assessment checklist, or arbitrating…
create-oss-skill
Create well-formed Agent Skills following the agentskills.io specification. Scaffold directories, write SKILL.md files, bundle scripts, and structure instructions for progressive disclosure. Use when creating a new skill, reviewing skill structure, optimizing a skill description, or setting up evals for skill quality.
extend-oss-skills-to-claude
Extend standard agentskills.io skills with Claude Code-specific features. Invocation control, subagent execution, dynamic context injection, string substitutions, model/effort overrides, and deployment scoping. Use when adapting a portable skill for Claude Code, adding Claude-specific frontmatter, setting up subagent…