writing-fragments
by mattpocockwriting-fragments is a focused interview skill for capturing claims, vignettes, sharp sentences, and half-formed ideas as raw material for a future article. It fits Knowledge Capture and early-stage drafting when you want to collect fragments first and defer structure until later. Use writing-fragments when you need a practical writing-fragments guide instead of a normal outline prompt.
This skill scores 67/100, which is enough to justify listing, but users should treat it as a specialized, moderately documented workflow rather than a turnkey package. The repository gives a clear use case and a concrete interaction model for collecting writing fragments into a single markdown file, so an agent can likely trigger and use it with less guesswork than a generic prompt. However, the install decision is only partially supported because the repo lacks supporting files, examples, and a visible install command, so adopters should expect to infer some operational details from the SKILL.md alone.
- Clear trigger and use case: the description says to use it when the user wants to develop ideas before structure, or mentions fragments/ideate/raw material.
- Operational workflow is explicit: interview the user relentlessly, preserve edits by re-reading the file, and append fragments to one markdown document.
- Good directional constraints: it forbids outlines/phases and defines what counts as a fragment, which reduces agent guesswork.
- Documentation is somewhat thin beyond the core flow: no supporting files, scripts, references, or resources are present.
- No install command is shown in SKILL.md, so users may need to figure out setup and invocation details themselves.
Overview of writing-fragments skill
The writing-fragments skill is for building an article from raw material instead of starting with an outline. It helps you run a focused interview, collect usable lines, and keep appending them into one markdown document as the draft takes shape. This is the right writing-fragments skill if you want to capture claims, vignettes, turns of phrase, and half-formed ideas before deciding what the final structure should be.
It is best for Knowledge Capture and early-stage drafting when the main problem is not polishing, but finding what is actually worth keeping. The core differentiator is that writing-fragments avoids forcing phases, sections, or a clean thesis too early. If you already know the structure, a normal prompt is usually enough; if you need raw material first, this skill is a better fit.
What this skill is for
Use writing-fragments when the user says they want to ideate, dump thoughts, gather fragments, or turn a rough subject into writing material. It is especially useful when the source content is messy or the writer is unsure what the final article should become.
What makes it different
The workflow is conversational and accumulative: ask, capture, reread, append. The skill expects the document to remain editable during the session, so preserving user edits matters as much as producing new text. That makes it more practical than a one-shot brainstorming prompt for live writing sessions.
When not to use it
Do not use writing-fragments if the user wants a finished outline, a structured brief, or a polished article from the start. It also is not the right fit when the output must be self-contained and cold-reader friendly immediately.
How to Use writing-fragments skill
Install and load context
Use the writing-fragments install flow from your skills directory, then open skills/in-progress/writing-fragments/SKILL.md first. In this repo, there are no helper scripts or support folders, so the skill behavior lives almost entirely in the skill file itself. That means install success depends on reading the instructions closely, not on discovering hidden automation.
Turn a rough goal into a good prompt
The skill works best when the user gives a topic, audience, and the kind of fragments they want to preserve. Strong inputs look like: “I’m writing about why small teams adopt AI note-taking tools; interview me for sharp claims and examples,” or “Help me capture fragments for a Knowledge Capture essay about why my system failed.” Weak inputs like “write something about productivity” leave too much ambiguity for the grilling session to produce useful material.
Practical workflow during the session
Start by asking where to save the file if no path was provided, then keep that path consistent for the rest of the session. Re-read the markdown file before every append so user edits are not overwritten. On the first write, keep only a single H1 working title at the top, then add fragments directly as they emerge rather than converting them into sections too early.
Best reading order in the repo
Begin with SKILL.md, then inspect the top-level guidance inside the same file for the definition of a fragment, the session style, and the file-format rules. Those are the parts that affect output quality most. If you are adapting the writing-fragments guide for your own workflow, focus on the constraints around structure, document state, and what counts as a fragment.
writing-fragments skill FAQ
Is writing-fragments good for beginners?
Yes, if the user can answer questions about a topic and tolerate an iterative process. It is less about writing skill and more about willingness to react to prompts and keep a document open while the session runs.
How is this different from a normal prompt?
A normal prompt often tries to generate a complete piece. writing-fragments is designed to collect material first and defer organization. That difference matters when the goal is to discover the article’s best lines before deciding its shape.
Does it fit Knowledge Capture workflows?
Yes. The writing-fragments for Knowledge Capture use case is strong when the goal is to preserve lived experience, insight, or interview responses as reusable writing fragments rather than forcing an immediate summary.
What should make me avoid it?
Skip it if you need structure, a final argument, or a single-pass deliverable. It is also a poor fit if the user cannot keep editing the same file, because the skill depends on preserving and rereading the current document.
How to Improve writing-fragments skill
Give the model better raw material
The best way to improve writing-fragments usage is to answer with concrete examples, contrasts, quotes, and specific moments instead of abstract themes. If you say “I need fragments about remote work,” you will get generic material; if you say “I want fragments about the moment a team stops trusting meetings,” you will get sharper writing fuel.
Protect the session from structure creep
A common failure mode is drifting into outlines too early. If the conversation starts summarizing instead of harvesting, redirect toward lines worth keeping, memorable claims, and details that could survive in the final article. The skill works better when you treat structure as a later editing step.
Iterate on what survives
After the first pass, review the document for fragments that feel readable, specific, and reusable. Ask for more material around the strongest lines, or ask the model to produce alternatives only for the weakest pieces. That is usually more effective than asking for a full rewrite, because the goal is to expand the fragment bank before synthesis.