story-context
Context scoping for writing subagents: use when deciding what context a subagent should receive, whether ephemeral story decisions should be written down before handoff, and how much to pass. Poor context handoffs cause writers to invent contradictions and critics to miss relevant history.
What this skill does
# Story Context Every subagent handoff starts with a context decision. Get it wrong and the writer invents facts that contradict established canon, the critic misses a continuity issue because it never saw the relevant chapter, or the brainstormer explores territory the author already rejected. A subagent only knows what you give it. This skill is the judgment: what story context to pass, when to write decisions down before handing off, and how much is enough. ## Choose the Right Mechanism Three ways to get context to a subagent, each for a different situation. **Point at files**: when context already exists as files — chapters, outlines, wiki pages, style files, character state. Default choice, because files are stable, inspectable, and survive compaction. Name the specific files in the subagent's prompt and scope tightly: pass the files that matter, not everything. **Put it in the prompt**: when the subagent needs decisions, reasoning, or brainstorm context that hasn't been written down yet. Capture the *why* behind choices directly in the prompt — why the author picked this angle, what they rejected. **Materialize first**: when context is too important to be ephemeral. If critical story decisions only live in the conversation, write them to `kb/` or `work/` *before* handing off. If a writer could accidentally contradict this context, materialize it. If it's supplementary background, putting it in the prompt is fine. ## What Each Agent Needs ### Writers Writers need enough to stay in voice and on-canon, not everything ever written: - **Scene brief or outline**: what happens in this scene, the beats to hit - **Relevant style files**: check `kb/styles/` and pick the files that match. Character files for whoever appears, scene-type files for the kind of scene. Each style file is self-describing: read the top to know when it applies. - **Continuity anchors**: the immediately preceding chapter or scene (for flow), plus any chapters that establish facts this scene references. Two to four files, not the entire manuscript. - **Character state**: character files from `kb/characters/` for characters who appear, especially if their emotional state or knowledge has changed recently - **Vocab**: relevant `vocab.md` files when the scene uses invented terms, magic/faction names, titles, relationship labels, or genre terms with project-specific meanings Tell the writer where to find more if it needs to explore, for example: "the full arc outline is in `work/outline/`; focus on the Route 1 section." Avoid passing everything preemptively. ### Critics Critics need the draft plus enough context to judge it against: - **The draft being reviewed** - **The scene brief or outline**: so the critic can check whether the draft achieved what it was supposed to - **Relevant style files**: so voice critics can compare against the target - **Prior chapters for continuity**: so continuity critics can cross-reference - **Author intent**: the direction the author discussed, via materialized decision notes or directly in the prompt - **Known issues**: tracked issues in `kb/issues/` if the critic should watch for specific recurring problems - **Vocab**: relevant `vocab.md` files when consistency of naming, aliases, deprecated terms, or invented language matters ### Brainstormers Brainstormers need constraints, not answers: - **The question being explored**: scoped tightly in the prompt - **Established context that constrains the answer**: character profiles, timeline, prior decisions that limit the design space - **What's been rejected**: so they don't re-propose dead ends - **Existing vocabulary**: enough vocab context to avoid minting new names for concepts the project has already named Don't pass too much: brainstormers that receive the full project history tend to produce conservative ideas that fit neatly into existing patterns instead of exploring fresh territory. ### Knowledge Maintenance - **Chronicler**: the chapter(s) to extract facts from, plus existing canon files, timeline entries, and vocab files for deduplication - **Continuity-checker**: the draft plus canon, timeline, character state, and vocab files for any domains the draft touches ## Cross-Phase Context Carry forward what a previous phase learned. The revision writer benefits from seeing what the first-draft writer discovered; the critic benefits from seeing prior critique rounds. When you run a multi-step loop, include the earlier phase's notes and the files it produced in the next subagent's prompt — reasoning context for the *why*, files for the *what*. ## Vocab Handoffs Treat vocabulary as operational story context. If a writer, critic, or brainstormer could choose the wrong name for a concept, pass the relevant `vocab.md` file or write the decision down before handing off. This matters most for magic systems, factions, recurring in-world phrases, titles, relationship labels, and terms the author corrected during conversation. When a session settles terminology, record it before handoff: canonical name, meaning, aliases still in circulation, and boundaries that prevent likely confusion.
Related in Writing & Docs
jax-development
IncludedUse this skill when the user is writing, debugging, profiling, refactoring, reviewing, benchmarking, parallelising, exporting, or explaining JAX code, or when they mention JAX, jax.numpy, jit, grad, value_and_grad, vmap, scan, lax, random keys, pytrees, jax.Array, sharding, Mesh, PartitionSpec, NamedSharding, pmap, shard_map, Pallas, XLA, StableHLO, checkify, profiler, or the JAX repo. It helps turn NumPy or PyTorch-style code into pure functional JAX, fix tracer/control-flow/shape/PRNG bugs, remove recompiles and host-device syncs, choose transforms and sharding strategies, inspect jaxpr/lowering/IR, and benchmark compiled code correctly.
nature-article-writer
IncludedDrafts, rewrites, diagnostically critiques, and style-calibrates primary research manuscripts for Nature and Nature Portfolio journals. Use when the user wants a Nature-style title, summary paragraph or abstract, introduction, results, discussion, methods, figure legends, presubmission enquiry, cover letter, reviewer response, or when a scientific draft sounds generic, jargon-heavy, structurally weak, or AI-ish and needs precise, broad-reader-friendly prose without inventing data, analyses, or references. Best for primary research articles and letters rather than reviews or press releases unless explicitly adapting one.
deckrd
IncludedDocument-driven framework that derives requirements, specifications, implementation plans, and executable tasks from goals through structured AI dialogue. Use when user says "write requirements", "create spec", "plan implementation", "derive tasks", "structure this feature", "break down into tasks", or "document this module". Also use for reverse engineering existing code into docs (/deckrd rev). Do NOT use for direct code writing — use /deckrd-coder after tasks are generated. Do NOT use when the user only wants to run or fix existing code without planning.
clinical-decision-support
IncludedGenerate professional clinical decision support (CDS) documents for pharmaceutical and clinical research settings, including patient cohort analyses (biomarker-stratified with outcomes) and treatment recommendation reports (evidence-based guidelines with decision algorithms). Supports GRADE evidence grading, statistical analysis (hazard ratios, survival curves, waterfall plots), biomarker integration, and regulatory compliance. Outputs publication-ready LaTeX/PDF format optimized for drug development, clinical research, and evidence synthesis.
handling-sf-data
IncludedSalesforce data operations with 130-point scoring. Use this skill to create, update, delete, bulk import/export, generate test data, and clean up org records using sf CLI and anonymous Apex. TRIGGER when: user creates test data, performs bulk import/export, uses sf data CLI commands, needs data factory patterns for Apex tests, or needs to seed/clean records in a Salesforce org. DO NOT TRIGGER when: SOQL query writing only (use querying-soql), Apex test execution (use running-apex-tests), or metadata deployment (use deploying-metadata).
accelint-ac-to-playwright
IncludedConvert and validate acceptance criteria for Playwright test automation. Use when user asks to (1) review/evaluate/check if AC are ready for automation, (2) assess if AC can be converted as-is, (3) validate AC quality for Playwright, (4) turn AC into tests, (5) generate tests from acceptance criteria, (6) convert .md bullets or .feature Gherkin files to Playwright specs, (7) create test automation from requirements. Handles both bullet-style markdown and Gherkin syntax with JSON test plan generation and validation.