Menu
Use when writing or fixing Python for run_python — browser WASM Pyodide contract, forbidden APIs, package substitutes, and webagent.http.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Use when shell commands fail, npx/curl/git won't run, or you must choose run_shell vs web_fetch vs file tools in Nodebox vs host.
Use when the user says remember this, save a preference or API key, or you must pick memory_save vs session notes vs skill (action=manage) vs wiki_* tools.
Use when the input is an image, screenshot, diagram, PDF/DOCX, audio file, or YouTube link—extract metadata/text before reasoning.
Use when the user wants to discover, search, or install agent skills from online registries—top results by installs, stars, or votes. Or user sends /find_skills.
Use when calling REST or GraphQL APIs with web_fetch/web_post — auth headers, query params, GraphQL shape, Directus GraphQL mutations/relations, discovery order, and reading error bodies.
Use when the user or agent installs a skill from skills.sh or GitHub — map WebFetch, Bash, Python, Playwright, and MCP references to Web Agent built-ins.
| name | Pyodide Runtime |
| description | Use when writing or fixing Python for run_python — browser WASM Pyodide contract, forbidden APIs, package substitutes, and webagent.http. |
| version | 1.0.0 |
| category | bundled |
| primary-tools | ["run_python","web_fetch","web_post","web_upload","skill"] |
| tags | ["pyodide","python","wasm","browser","run_python","micropip","webagent.http"] |
| triggers | ["run_python","pyodide","python script","micropip","browser python","wasm python","python in browser","incompatible python"] |
| Need | Use |
|---|---|
| Run Python in workspace | run_python (code or path, optional cwd, env, packages, timeout_ms) |
| One-off REST from agent | web_fetch / web_post / web_upload — not requests in chat args |
HTTP inside a reusable .py | import webagent.http as http (proxy-backed) |
| File bytes to API | http.upload_file in script, or agent web_upload |
| Shell / git / curl / npx | browser-runtime-map — not available in Nodebox |
| Import map & blocks | Static preflight before worker; see browser-runtime-map Pyodide table |
Non-negotiable: Read this skill (or skill (action=view) browser-runtime-map) before generating Python. Capability manifest: src/runtimes/webcontainer/pyodide-capabilities.json. Substitution patterns: src/runtimes/webcontainer/pyodide-substitution-matrix.json.
run_python call or skill script that will execute in the browser.pyodide_* errors.Not for: simple PDF/DOCX text extraction — try pdf_extract / docx_extract first.
Your execution environment is not native Python.
| Fact | Implication |
|---|---|
| Pyodide + WASM in a browser tab | No OS, no Docker, no real /home/user |
| Virtual Emscripten FS | Write under run_python cwd; host mirrors workspace |
| No subprocess / native binaries | No soffice, git, ffmpeg, pdftoppm, tesseract, … |
| No raw TCP | No psycopg2, pymongo, redis, socket.create_connection |
| Networking = browser fetch | CORS + /api/proxy; use webagent.http or agent HTTP tools |
| Threads limited | Prefer async + await asyncio.sleep(0) for long work |
| Packages | Stdlib, Pyodide wheels (auto-loaded), micropip allowlist only |
subprocess, multiprocessing, os.system, pty, pexpect, signal, resource, pwd, grp, fcntl, asyncio.create_subprocess_exec, playwright, selenium, torch, tensorflow, cv2, pdf2image, host DB drivers opening TCP.
threading for parallelism, raw socket clients, requests/httpx for new code (use webagent.http), sys.stdout.reconfigure, LD_PRELOAD / ctypes.CDLL("*.so"), unbounded sync loops.
web_fetch / web_post / web_uploadimport webagent.http as httpjson, csv, pathlib, zipfile, re, hashlib, …pypdf, python-docx, openpyxl, Pillow, numpy, … (often auto-loaded)pyodide.http.pyfetch only when same-origin/CORS allows direct fetchIf any answer is wrong → redesign (see substitution matrix).
# Good — inside run_python
import webagent.http as http
resp = http.get("https://api.example.com/data")
data = resp.json()
# Bad — host assumptions
import requests
requests.get("http://localhost:3000")
browser-runtime-maphttp-apiimported-skill-compat (script_warnings from preflight)zipfile here; extract: extract_archiverun_shell python3 as full CPython — use run_python with Pyodide rules.pip install in script — use packages arg or preflight auto-load / micropip allowlist.~/.config, /var/tmp, or persistent paths outside workspace cwd.web_upload / http.upload_file.subprocess.run(["git", …]) or os.system("curl …") in browser Python.import playwright / from pdf2image import convert_from_path after preflight already warned.while True: pass without timeout_ms on the tool call.