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 Reederey87/grok-build-codex-plugin --skill grok-build-delegategit clone --depth 1 https://github.com/Reederey87/grok-build-codex-pluginWrote 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/reederey87/grok-build-codex-plugin/grok-build-delegate)<a href="https://agentmods.dev/skills/reederey87/grok-build-codex-plugin/grok-build-delegate"><img src="https://agentmods.dev/badge/skills/reederey87/grok-build-codex-plugin/grok-build-delegate/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/reederey87/grok-build-codex-plugin/grok-build-delegate"><img src="https://agentmods.dev/badge/skills/reederey87/grok-build-codex-plugin/grok-build-delegate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00032 | $0.00914 |
| Opus 5 | $0.00016 | $0.00457 |
| Sonnet 5 | $0.00006 | $0.00183 |
| Haiku 4.5 | $0.00003 | $0.00091 |
Grade A, and why
grok-build-delegate 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 8d 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Grok Build delegate
Bridge to the local, subscription-authenticated Grok Build CLI (xAI). Use it to get an independent second opinion, an adversarial review, or a delegated investigation without leaving the Codex session. This is a separate review/task channel, not a model swap for your own work.
When to use
Reach for these tools only when the user explicitly asks for Grok, Grok Build, grok-4.5, a second opinion, or an external/adversarial review. Do not route ordinary work here.
Routing
- Setup or health check →
grok_build_setup. Run it first when anything looks misconfigured; it reports CLI availability, auth, model, fallback, and state storage. Userun_smoke_test: truefor one model orsmoke_models: "all"only when the user asks to verify every discovered model. - Quick question or second opinion →
grok_build_prompt. Foreground, strictly read-only. - Code review →
grok_build_review. Preview withpreview_only: true, explain that path-filtered, best-effort content-redacted repository context will be sent to xAI, and execute only withexternal_disclosure_acknowledged: true. Passscope(working-tree,branch, orrepo) andbase_refforbranch; setadversarial: trueto challenge assumptions. Passmodelwhen the user requests a specific Grok model. - Investigation or a longer task →
grok_build_task. For long runs setbackground: true, then pollgrok_build_statusand readgrok_build_resultwhen it finishes. Usegrok_build_cancelto stop a running job.
Every tool takes cwd — pass the absolute path of the repository being worked on.
Safety rules
- Default to read-only.
grok_build_promptandgrok_build_reviewcan never write. Their local Grok invocations use thestrictsandbox and expose onlyRead,Grep, andGlob, not Bash. - Writes go through
grok_build_taskwithwrite: true, and only when the user explicitly asks Grok to change files. This is double-gated: the server also requiresGROK_BUILD_ALLOW_WRITE=truein its environment. Without it the tool returnswrite mode disabled — set GROK_BUILD_ALLOW_WRITE=true to enable. Relay that to the user rather than trying to work around it. - Never set
always_approveunless the user explicitly says "yolo" or "always approve". It only applies to write tasks and requiresGROK_BUILD_ALLOW_ALWAYS_APPROVE=true; the bridge keeps Grok inside its curated capability allowlist rather than forwarding a global CLI bypass. - Never put secrets, credentials, tokens, or env-file contents in a prompt.
- Grok returns findings and proposed patches. Apply any change through normal Codex edits under Codex's own sandbox and approval policy — do not treat a Grok proposal as approved.
- On an auth failure, tell the user to run
grok login(orgrok login --device-auth).
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.
- 8d ago First seen · 67 lines · 32 tokens per session scan A 127440e12e8d
grok-build-delegate is a skill published in the GitHub repository Reederey87/grok-build-codex-plugin (4 stars, last pushed 5d ago), licensed Apache-2.0. It adds 32 tokens to every session and 914 once invoked, about $0.0002 per session on Opus 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 2026-08-31.
Other skills, from other repositories
code-review-excellence
This skill should be used when the user asks to review a diff or pull request, write review comments, audit code quality, establish review standards, or improve how a team performs code review.
stop-that-shit
Keep coding agents focused on requested and necessary work. Use when a request sets a read-only, answer-only, file, action, or stopping boundary; when evidence shows scope creep, speculative hardening, unnecessary hashing or dependencies, repeated audit loops, or valueless delegation; or when the user invokes Stop…
stss
Reduce defensive disclaimers, stacked hedging, and self-protective narration in proposals and decision-facing writing. Use when the user asks to rewrite or audit a proposal, plan, research contribution, executive summary, or similar text for directness. Do not use for ordinary code work or unrelated prose.
do-it-review
Use to assess requirements and implementation quality independently, or to resolve a batch of review findings.
tfx-prune
A cleanup workflow that asks three AI coding assistants to independently identify unnecessary code, then removes only issues they agree are unwanted. It checks for duplication, needless abstractions, excessive error handling, and similar readability problems.
tfx-review
A code-review workflow that asks several command-line AI reviewers to inspect changes independently, then reports issues that at least two reviewers agree on. TDD means test-driven development, where tests are written to guide implementation, but this add-on is for reviewing code rather than defining TDD.