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 shinpr/linear-prism --skill recipe-decomposegit clone --depth 1 https://github.com/shinpr/linear-prismWrote 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/shinpr/linear-prism/recipe-decompose)<a href="https://agentmods.dev/skills/shinpr/linear-prism/recipe-decompose"><img src="https://agentmods.dev/badge/skills/shinpr/linear-prism/recipe-decompose.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 | $0.00057 | $0.00635 |
| Opus 5 | $0.00028 | $0.00318 |
| Sonnet 5 | $0.00011 | $0.00127 |
| Haiku 4.5 | $0.00006 | $0.00064 |
Grade A, and why
recipe-decompose 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 8d 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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Orchestrator Definition
I am an orchestrator. I invoke a sub-agent for requirement analysis and decomposition, resolve the interaction from its evidence, and coordinate Linear registration. All Linear MCP operations are executed directly by the orchestrator. See references/register.md for the registration procedure.
Sub-agent: linear-prism:task-decomposer-linear
Input
$ARGUMENTS contains one of:
- A Linear issue URL (e.g.,
https://linear.app/team/issue/ENG-123) - A file path to a PRD or requirement document
- Empty (ask the user which input to use)
Prerequisites
- Linear MCP connection is active. Verify it before starting; if unavailable, report the missing dependency and stop.
Workflow Overview
Round 1: Fetch input → Analyze outcome and scope → Continue or ask for a blocking user decision
Round 2: Propose decomposition strategy → Present task list → User adjusts
Round 3: Explicit approval of the current valid task list → Register in Linear
Detailed procedures for each round are in the references/ directory:
- Round 1: See
references/analyze.md - Round 2: See
references/plan.md - Round 3: See
references/register.md
Scope Boundaries
In scope:
- Fetching and analyzing requirement content from Linear or local files.
- Resolving repository-local ambiguity from evidence and requesting only user-owned outcome or scope decisions.
- Decomposing requirements into typed tasks with dependency graphs.
- Registering tasks as Linear issues.
Out of scope:
- Modifying the source PRD or Linear issue directly.
- Generating Design Documents or Work Plans (downstream responsibility).
- Implementing any task produced by this skill.
Error Handling
| Condition | Action |
|---|---|
| Linear MCP unavailable | Report the missing required dependency and stop |
| Input URL returns empty or inaccessible content | Present the error, ask user for alternative input |
Agent returns recommended_action: needs_input |
Ask only the returned user-owned questions, then re-analyze |
| User rejects task list in Round 2 | Collect adjustment instructions, re-invoke agent in revise mode (see references/plan.md) |
What ships with it
4 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.
- 8d ago First seen · 66 lines · 57 tokens per session scan A f5de17e11ea1
recipe-decompose is a skill published in the GitHub repository shinpr/linear-prism (7 stars, last pushed 19d ago), licensed MIT. It adds 57 tokens to every session and 635 once invoked, about $0.0003 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
storythis
Turn FINISHED design/spec work into the right Linear structure for the dev team — adaptively. Creates the subset that fits: a value-framed Project (epic) + a shared design issue + lean, code-verified build stories, OR just some of those. Right-sizes — not everything needs a project: one small change → one issue; a…
linearthis
Adaptive brainstorming partner that shapes work WITH the user into well-formed Linear structure — an epic + work issues, or a single standalone issue — with acceptance criteria, from a one-line idea, a messy backlog, or an existing epic to refine. Meets you where you are: pressure-tests a fuzzy idea, or just confirms…
ccthis
Build ONE Linear work issue end-to-end — a sub-issue of an epic, or a standalone issue: set it In Progress, build with gstack, keep Linear updated, and ship it as its own branch + PR. Given an epic ID, it picks ONE work issue to build this run and says which. It completes only the work issue it built and verified …
codex-symphony
Install and operate a portable OpenAI Symphony + Linear orchestration setup in any Git repository. Use when the user wants a one-command local Symphony runner, a reusable WORKFLOW template, background launch scripts, or a Codex wrapper that restarts Symphony automatically when reopening the CLI.
chonchi
Hand the CURRENT session's finished, user-tested work issue off to the broader team for review. Usually run bare — no ID needed, since you've been building one work issue on one branch all session. Moves that work issue to In Review, comments a summary with actual-vs-estimate, attaches the FULL session transcript as…
neat-freak
Knowledge and governance closeout: reconcile project docs, rule files (CLAUDE.md/AGENTS.md), authorized agent memory, and workspace residue with what the code and runtime actually do, so the next session or the next person starts from one current answer. Trigger when the user names "neat-freak", "洁癖", or "/neat" — and…