filesystem-context
An agent skill by muratcankoylan, from muratcankoylan/Agent-Skills-for-Context-Engineering. Tags: automation, cli, developer-tools, documents, local-workflow.
What it does
This skill should be used when agent work needs file-backed context: durable scratchpads, tool-output offloading, just-in-time discovery, cross-agent handoff files, filesystem memory, or cleanup policies for context stored outside the prompt.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add muratcankoylan/Agent-Skills-for-Context-Engineering --skill filesystem-contextOr 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/muratcankoylan/Agent-Skills-for-Context-Engineering
cp -r Agent-Skills-for-Context-Engineering/skills/filesystem-context ~/.claude/skills/filesystem-contextSafety box score
Grade B, low risk: 18 of 100 risk points, rated 2026-10-14. 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 local filesystem context manager: it writes scratch files, plan JSON, workspace logs, and can clean up demo artifacts. Its main risks are persistent file accumulation, a demo path that deletes its own scratch directory, and self-modifying instruction patterns that can steer future agent behavior from files.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 1 (low) | A native Python module is shipped and runnable via `if __name__ == "__main__":`, so it executes local demo code and file I/O, but it does not run shell commands or eval code. |
| Network access | 0 (none) | No code downloads from or uploads to remote hosts; the only remote items are named references in prose. |
| Filesystem reach | 1 (low) | The skill is built around local writes under `scratch/`, `demo_scratch/`, and similar working-tree paths, with cleanup scoped to those directories. |
| Secrets & credentials | 0 (none) | There is no access to token stores, key files, `.env` files, browser sessions, or pasted credentials. |
| Privilege & persistence | 0 (none) | No sudo, global installs, service registration, shell-profile edits, or other privilege-changing operations appear. |
| Instruction hijack surface | 2 (moderate) | The docs explicitly encourage loading instructions from files and even persisting learned preferences back into instruction files, e.g. `write learned preferences and patterns to their own instruction files`, which can reshape future behavior from file contents. |
| Irreversible actions | 0 (none) | It does not push, post, deploy, pay, book, or delete external resources; actions are limited to local scratch/demo cleanup. |
| Transparency | 0 (none) | All shipped content is plain text source and markdown; there are no binaries, obfuscated blobs, or runtime-downloaded installers. |
Before installing
- Read `scripts/filesystem_context.py` and note the `shutil.rmtree(demo_path)` cleanup path.
- Decide whether you want the self-modifying preference pattern described in `SKILL.md` and `references/implementation-patterns.md`.
- If you test it, run it in a disposable checkout because it creates `scratch/` and `demo_scratch/` files.
Rated 2026-10-14 by an automated review of 3 files from the skill's repository.
Source
- Repository
- muratcankoylan/Agent-Skills-for-Context-Engineering (all skills from this repository)
- Path
- skills/filesystem-context/SKILL.md
- Branch
- main
- Updated
- 2026-09-19
Related skills
- book-to-skill — Converts books and documents (PDF, EPUB, DOCX, HTML, Markdown, plain text, RTF, MOBI/AZW with Calibre) into structured agent skills, extracting frameworks.
- dxf — Generate, regenerate, and validate 2D DXF drawings from Python build123d sources.
- agents-get-started — Use when a developer wants to create a new agent project or get started with AgentCore.
- cli-for-agents — Designs or reviews CLIs so coding agents can run them reliably: non-interactive flags, layered --help with examples, stdin/pipelines, fast actionable errors.
- devcontainer-setup — Creates devcontainers with Claude Code, language-specific tooling (Python/Node/Rust/Go), and persistent volumes.
- modern-python — Configures Python projects with modern tooling (uv, ruff, ty). Use when creating projects, writing standalone scripts, or migrating from pip/Poetry/mypy/black.