Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add decodingai-magazine/building-a-coding-agent-from-scratch-course --skill demo-5-sandbox-feature-prgit clone --depth 1 https://github.com/decodingai-magazine/building-a-coding-agent-from-scratch-courseWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-5-sandbox-feature-pr)<a href="https://agentmods.dev/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-5-sandbox-feature-pr"><img src="https://agentmods.dev/badge/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-5-sandbox-feature-pr/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-5-sandbox-feature-pr"><img src="https://agentmods.dev/badge/skills/decodingai-magazine/building-a-coding-agent-from-scratch-course/demo-5-sandbox-feature-pr.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Tool Misuse · line 74 Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).Fix: Validate all tool parameters against an allowlist. Reject dangerous parameter values (shell=True, --force, -rf /) and use safe defaults.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00060 | $0.01247 |
| Opus 5.5 | $0.00024 | $0.00499 |
| Sonnet 5.5 | $0.00012 | $0.00249 |
| Haiku 4.5 | $0.00006 | $0.00125 |
Grade A, and why
demo-5-sandbox-feature-pr scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The meta demo: decode improves decode. Run decode inside an isolated sandbox Workspace that is a clone of this very course repo, plan and implement one small self-contained feature there, and let the harness ship the work back as a branch you turn into a draft PR.
Nothing the model does touches your host checkout — the Workspace is an isolated clone, and the branch only lands on your host through Hand-back (ADR-0012 §8).
1. Launch decode against a sandboxed clone of the course repo (human step)
Sandbox mode is selected by the SANDBOX_MODE environment variable — it is an env var, not a
command-line flag; the repo to clone into the Workspace is the --repo flag. Launch the local
docker rung (Docker must be running):
SANDBOX_MODE=docker decode --repo [email protected]:decodingai-magazine/building-a-coding-agent-from-scratch-course.git
This clones the course repo into the isolated Workspace (/workspace ≡ host .decode/sandbox)
and drops you into the REPL, with bash and the file tools scoped to that clone.
To run the same demo on the remote rung instead, swap SANDBOX_MODE=docker for
SANDBOX_MODE=modal — everything else is identical (Modal credentials must be configured).
2. Plan the feature — in plan mode
If you are decode reading this inside the Workspace, start here — step 1 is the human's launch
command, already done; do NOT try to run decode yourself. Enter plan mode by calling
enter_plan_mode with no arguments (it takes none — you present the plan later, to
exit_plan_mode), explore the code read-only, then present a short plan before touching anything.
Pick ONE feature — small, self-contained, provable in a single session:
- A
decode --versionflag that prints the installed package version, plus its unit test. - One small pure helper (e.g. in
tools/orentities/) with a focused unit test. - A scoped docstring/README fix that resolves a real inaccuracy you found while exploring.
The plan should name the files to touch, the test that proves the feature, and nothing else.
Exit plan mode (exit_plan_mode) once the human approves it.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 102 lines · 60 tokens per session scan A b163c95c8688
demo-5-sandbox-feature-pr is a skill published in the GitHub repository decodingai-magazine/building-a-coding-agent-from-scratch-course (1,034 stars, last pushed 5d ago), licensed Apache-2.0. It adds 60 tokens to every session and 1,247 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
repo-hygiene
Use when the scheduled repo-hygiene workflow runs from GitHub Actions (or an operator dry-run) to scan the repository for small, certain docs/test/code hygiene issues and fix them as one batched branch.
contributor-onboarding
Help a new contributor get productive on this checkout - inspect sync state against main, build, run the repository's exact verification gate, and produce a local what's-new digest. Never fetches, pulls, or modifies a dirty tree on its own. Explicit-only.
codew-release-qa-sweep
Use before claiming Codewhale release work is done: run the full gate sweep and list the manual QA targets.
gh-find-prs
Survey open Codewhale PRs and triage each for mergeability and disposition against the real landing branch.
codex-pr-body
Update the title and body of one or more pull requests.
update-v8-version
Update Codex's pinned v8 / rustyv8 versions, validate the release-candidate path, and investigate failed V8 canary or artifact builds. Use when asked to bump V8, update rustyv8 artifacts, prepare or validate a V8 release candidate, check v8-canary, or diagnose why a V8 version update no longer builds.