writing-fragments
An agent skill by mattpocock, from mattpocock/skills. Tags: analysis, brainstorming, content-brief, documentation, writing.
What it does
Writing, explore: mine raw fragments, no structure yet.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add mattpocock/skills --skill writing-fragmentsOr copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/mattpocock/skills
cp -r skills/skills/in-progress/writing-fragments ~/.claude/skills/writing-fragmentsSafety box score
Grade B, low risk: 23 of 100 risk points, rated 2026-10-20. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This skill is a lightweight writing aid with no code execution or network activity, but it does persist conversation fragments into a markdown file. The main risks are file writes to a user-chosen path, including the initial prompt, and instructions that reduce per-write confirmation.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 0 (none) | No shell commands, scripts, or code evaluation are present; the skill is pure prompt text and file-writing guidance. |
| Network access | 0 (none) | The files contain no download, upload, API, or remote-host communication behavior. |
| Filesystem reach | 2 (moderate) | It appends to "a single markdown file", asks where to save if no path is passed, and says "Before every write: re-read the file from disk" plus "never overwrite the file; only append (or... edit a specific fragment in place)." |
| Secrets & credentials | 1 (low) | It will persist whatever the user says, including "Capture fragments from the very first thing the user says, including the initial prompt," so pasted secrets could end up saved to disk. |
| Privilege & persistence | 0 (none) | There is no request for sudo, global installs, services, shell startup edits, or other privilege-changing actions. |
| Instruction hijack surface | 2 (moderate) | The skill pressures the agent to skip routine confirmation with "Append silently. Don't ask permission for each fragment," which is a workflow-level instruction to reduce user check-ins. |
| Irreversible actions | 0 (none) | It does not post, send, deploy, book, pay, or otherwise act beyond writing a local markdown file for the user. |
| Transparency | 0 (none) | All content is readable markdown/YAML; there are no binaries, encoded blobs, or fetched installers. |
Before installing
- Check whether you are okay with the skill saving your prompts, including the initial one, into a markdown file.
- Use a scratch path first and confirm it only writes where you told it to write.
- Avoid using it in conversations containing secrets or sensitive pasted text unless you want them persisted.
Rated 2026-10-20 by an automated review of 2 files from the skill's repository.
Source
- Repository
- mattpocock/skills (all skills from this repository)
- Path
- skills/in-progress/writing-fragments/SKILL.md
- Branch
- main
- Updated
- 2026-10-05
Related skills
- teach — Explain a body of work plainly so a person actually understands it. Runs the `how` and `why` skills and weaves what they find into one clear explanation.
- workflow-from-chats — Extract durable working preferences from recent Cursor chats and convert them into skills, rules, or workflow docs.
- writing-beats — Writing, exploit; assemble raw material into a journey of beats, grounding each term before a beat leans on it.
- arena — Spawn N parallel candidates at the same task, pick a base, graft the strongest parts of the losers into it.
- audit-prep-assistant — Prepares codebases for security review using Trail of Bits' checklist.
- automate-me — Use for "automate me", "create/update/refresh my -mode skill", "turn/capture my preferences or working style into a skill".