Install
$ agentstack add skill-skryx-l-a-project-kit-content-writing ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
content-writing — books, courses, docs, and long-form prose
This sub-skill stands up a content / writing project: a substantial written work organised by outline and chapters, drafted iteratively, and exported to a polished format (PDF / DOCX / slides). It treats structure, voice, and audience-fit as first-class — the deliverable is prose, not code, so it uses an outline-and-chapter skeleton instead of source/build folders.
What this sub-skill is for
Long-form written deliverables: a book or e-book, an online course/curriculum, a documentation set or manual, a multi-part article or guide series, a newsletter run. Not for website copy that lives inside a site (website), not for a research report whose point is a decision (research-decision, though it composes when a book is research-heavy). Composes with website when the content also ships as a published site.
Mandatory grill-questions (fold into the Definition of Ready)
Lock these before drafting a single section:
- Format — book / e-book / course / docs / article series / manual. Each implies a different
structure (chapters vs. modules+lessons vs. reference pages). Recommend confirming the format before outlining.
- Audience — who reads it, their prior knowledge, what they should be able to do afterward.
Drives reading level, tone, and assumed background.
- Goal & key takeaways — what the reader/learner walks away with; the through-line of the work.
- Length & depth — target word/page count (or lesson count), and depth per unit. Sets the
outline's granularity.
- Outline — the chapter/module breakdown. Get the outline approved before any drafting —
this is the gate; drafting against an unapproved outline wastes work.
- Voice & language — tone (authoritative / friendly / academic), language/locale, and any style
guide or sample to match.
- Source material — does the user supply research/notes/transcripts, or do we research? Any
facts/claims that need verification or citation?
- Export targets — final formats (PDF, DOCX, EPUB, slide deck) and any layout/branding
requirements.
Project sub-agents to generate (into /.claude/agents/)
outliner(delegate-by-default) — turns the brief into a structured chapter/module
outline with per-unit goals and key points; iterates with the user until the outline is approved (the gate). Chains doc-coauthoring.
drafter(delegate-by-default) — writes each chapter/lesson to the approved outline in the
project's voice and target length; works unit-by-unit, never dumps the whole book at once.
editor— line-and-structure edits for clarity, consistency, voice, and flow; checks
continuity across chapters and flags claims needing sources.
- Plus the kit defaults:
reviewer(audits drafts against outline + audience + length) and
verifier (confirms the export actually builds and reads correctly in the target format).
Tools / CLIs / MCP / skills needed
Check in environment-readiness; offer install, never auto-install:
- Export tooling —
pdf(final PDF + assembly/merge),docx(manuscripts, TOC, headings,
letterhead), pptx (course/slide companion), xlsx only if the work needs data tables. pandoc for EPUB/format conversion if a book needs it.
- CHAIN these GLOBAL skills (the core of this type):
doc-coauthoring— the structured co-authoring workflow: context transfer, iterative
refinement, reader-verification. Drive outlining and drafting through it.
pdf/docx/pptx— produce the polished export(s).deep-research— when chapters need sourced facts; cite and mark verified vs. reasoned.verify— confirm the exported artifact opens and reads as intended.
File / asset nudges (on top of the base set)
Beyond CLAUDE.md, PROJEKT_.md, TASKS.md, DONE.md, README, .claude/ — use a manuscript skeleton, not a code tree:
OUTLINE.md— the approved chapter/module structure with per-unit goals (the source of truth).chapters/(ormodules/) — one Markdown file per chapter/lesson, drafted in order.STYLE.md— voice, tone, terminology, formatting, and language/locale conventions.research/+sources.md— supporting notes and the citation list, if the work is sourced.assets/— figures, diagrams, cover art;export/— the built PDF/DOCX/EPUB/deck outputs.front-matter/— title page, TOC, preface/intro, about — for book/manual formats.
Stack defaults & done-bar
Defaults: Markdown chapters under chapters/, an approved OUTLINE.md as source of truth, a STYLE.md voice guide, doc-coauthoring driving the write loop, exported to PDF/DOCX (EPUB for books) via the document skills. Draft unit-by-unit; never generate the whole work in one pass. Done-bar (all true): the outline is approved by the user; every chapter/lesson is drafted to its outlined goal and within target length; the editor pass is done (consistent voice, no continuity breaks, claims sourced or flagged); the work is exported cleanly to the target format(s) and the export opens and reads correctly.
Guardrails
- Outline approval is a gate — do not bulk-draft against an unapproved outline.
- Honesty in claims: any fact, statistic, quote, or citation must be real and verifiable — mark
verified vs. reasoned, and flag anything the user must confirm. Never invent sources or data.
- No plagiarism — write original prose; quote and attribute properly; study references for
patterns, don't lift text.
- Match the agreed voice and length — don't pad to hit a count or drift off the style guide;
flag if the target length doesn't fit the material.
- AI-writing honesty — if the user intends to publish, be candid about disclosure norms and
don't dress up thin material as authoritative.
- No emojis in the deliverable unless the user's style guide explicitly calls for them
(standing user rule: typographic symbols only by default).
- Commits under the user's own name only (Skryx-L-A); never add Claude as a co-author.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Skryx-L-A
- Source: Skryx-L-A/project-kit
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.