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/0xmariowu/autosearch/delegate-subtasknpx skills add 0xmariowu/Autosearch --skill delegate-subtaskgit clone --depth 1 https://github.com/0xmariowu/AutosearchWrote 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/0xmariowu/autosearch/delegate-subtask)<a href="https://agentmods.dev/skills/0xmariowu/autosearch/delegate-subtask"><img src="https://agentmods.dev/badge/skills/0xmariowu/autosearch/delegate-subtask.svg" alt="Measured on agentmods" 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 | $0.00071 | $0.01231 |
| Opus 5 | $0.00036 | $0.00616 |
| Sonnet 5 | $0.00014 | $0.00246 |
| Haiku 4.5 | $0.00007 | $0.00123 |
Grade A, and why
autosearch:delegate-subtask 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 5d 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 — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Delegate Subtask — Execution Contract
decompose-task splits a problem into sub-questions. This skill says how to execute each sub-question with a stable, auditable contract: inputs, budget, outputs, failure modes. Borrowed from MiroThinker + DeepAgents + deer-flow + DeepResearchAgent subagent patterns.
Contract
input:
id: str # stable subtask id, e.g. "sub_1" / "sub_1a"
parent_id: str | null # linking back to the decompose-task output
question: str # one specific sub-question
rationale: str # why this subtask matters for the parent goal
scope: list[str] # channels / tools the subtask may touch
budget:
latency_seconds: int
cost_usd: float
tool_calls: int # max total tool invocations
context_seed: list[dict] # evidence already gathered the subtask should start with
stop_conditions: list[str] # e.g. "answer rubrics satisfied" / "budget exhausted"
output:
id: str # echoes input.id
status: "success" | "partial" | "failure"
summary: str # 3-6 sentences; what was found
evidence: list[dict] # slim-dict Evidence items the subtask produced
citations: list[str] # URL list, matched to evidence
follow_ups: list[str] # open questions, if partial
metrics:
latency_ms: int
cost_usd: float
tool_calls: int
channels_hit: list[str]
failure_reason: str | null
Invocation Policy
- One subtask per thread/session — isolation matters. Do not merge two subtasks' tool calls into one session.
- Budget is the governor. Subtask must halt when ANY budget axis is exhausted and report
status: "partial". - Read
context_seed, don't re-search it. Seed is evidence the parent already has; subtask should build on, not duplicate. - Return slim evidence — use autosearch's
Evidence.to_slim_dict()shape so the parent can dedupe/merge. - Follow-ups are first-class. If a subtask runs out of budget but finds a promising lead, emit that in
follow_upsfor the parent planner to decide.
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.
- 5d ago First seen · 110 lines · 71 tokens per session scan A e5b9ea007969
autosearch:delegate-subtask is a skill published in the GitHub repository 0xmariowu/Autosearch (44 stars, last pushed 29d ago), licensed MIT. It adds 71 tokens to every session and 1,231 once invoked, about $0.0004 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.
Other skills, from other repositories
sanity-best-practices
Sanity development best practices for schema design, GROQ queries, TypeGen, Visual Editing, images, Portable Text, Studio structure, localization, migrations, Sanity Functions, webhooks, Blueprints, and framework integrations such as Next.js, Nuxt, Astro, Remix, SvelteKit, Angular, Hydrogen, and the App SDK. Use this…
portable-text-serialization
Render and serialize Portable Text to React, Svelte, Vue, Astro, HTML, Markdown, and plain text. Use when implementing Portable Text rendering in any frontend framework, building custom serializers for non-standard block types, converting Portable Text to HTML strings server-side, converting Portable Text to Markdown…
video-perception
Use when the user mentions a video file (.mp4, .mov, .avi, .mkv, .webm), a YouTube URL, asks to watch/analyze/review a video, or references video content in conversation.
prior-art-search
Systematic 7-step methodology for comprehensive patent prior art searches and patentability assessments using BigQuery and CPC classification.
bigquery-patent-search
Fast, cloud-based patent searching across 100 million+ worldwide patents using Google BigQuery - keyword search, CPC classification, patent details retrieval.
development-assistant
Guides through adding new features, MCP tools, analyzers, and extending the patent creator system.