CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

grok-build

Operate the Grok Build CLI (`grok`), including interactive and headless runs, authentication, sessions, worktrees, permissions, sandboxing, MCP servers, plugins, inspection, updates, and automation output. Load only when the user explicitly authorizes this skill or names Grok Build / grok-build. Before any command, confirm the `grok` binary is available; if missing, stop and give install or auth guidance.

Quality

95%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

100%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is an exemplary CLI-operation skill: lean, command-first, and version-aware (deferring to installed help and official docs), with clear sequencing, explicit validation and destructive-operation checkpoints, and a clean split between the overview and two real one-level-deep reference files. No dimension shows meaningful weakness.

DimensionReasoningScore

Conciseness

The body is ~130 lines of terse directives and command blocks with zero concept explanations Claude already knows — even the safety framing is compressed ("Do not describe `--always-approve` as equivalent to sandboxing") — so every token earns its place (anchor 5), not anchor 4 since no padding is trimmable.

5 / 5

Actionability

Fully executable, copy-paste-ready commands cover the common cases: `command -v grok`, the complete headless invocation with `--no-auto-update --cwd --sandbox -p ... --output-format json`, `grok --worktree --ref main "Fix the flaky test"`, and `grok worktree rm --dry-run <id>` — a clear anchor 5 match.

5 / 5

Workflow Clarity

The evidence-first sequence (1-4: verify binary, inspect help, inspect config, defer to installed docs) includes explicit validation and a failure feedback loop ("if missing, stop and give install or auth guidance"), plus checkpoints for destructive operations ("Preview removal with `grok worktree rm --dry-run <id>` before deleting", "Verify destructive commands such as session deletion... before running them"), so the destructive-operation cap does not apply — anchor 5.

5 / 5

Progressive Disclosure

The body is a genuine overview with two well-signaled, purpose-stated, one-level-deep references ("Read [references/cli-reference.md](references/cli-reference.md) for the subcommand and common flag index", "Read [references/workflows.md](references/workflows.md) for headless execution, sessions, worktrees..."); both files exist and nest no further — anchor 5.

5 / 5

Total

20

/

20

Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: it states a concrete capability scope, gives an explicit and restrictive load condition with concrete trigger names, and adds operational pre-flight guidance. The only slack is that capabilities are expressed as one verb over a domain list rather than multiple distinct actions, and a few adjacent natural keywords (e.g., xAI) are missing.

Suggestions

Lead with 2-3 distinct concrete action verbs (e.g., 'Run grok interactively or headlessly, manage sessions and worktrees, configure permissions, sandboxing, MCP servers, and plugins') instead of a single 'Operate' verb followed by a domain list.

Add one or two adjacent natural trigger phrasings users might say, such as 'xAI CLI' or 'run grok in CI', to broaden keyword coverage without adding vagueness.

DimensionReasoningScore

Specificity

Names the concrete top-level action "Operate the Grok Build CLI (`grok`)" and comprehensively enumerates its scope ("interactive and headless runs, authentication, sessions, worktrees, permissions, sandboxing, MCP servers, plugins, inspection, updates, and automation output"), but relies on a single verb plus domain areas rather than multiple distinct parallel actions, so it sits at anchor 4 rather than 5.

4 / 5

Completeness

Explicitly answers both: what ("Operate the Grok Build CLI (`grok`), including interactive and headless runs, authentication, sessions...") and when ("Load only when the user explicitly authorizes this skill or names Grok Build / grok-build"), with concrete trigger conditions and a pre-flight rule ("confirm the `grok` binary is available; if missing, stop and give install or auth guidance") — a clear anchor 5 match, not 4, since the 'when' is fully explicit.

5 / 5

Trigger Term Quality

Covers the natural name variants users would say — "Grok Build CLI (`grok`)", "names Grok Build / grok-build", "the `grok` binary" — plus domain terms like "headless" and "automation", but omits adjacent natural phrasings such as the vendor name ("xAI") or "run grok", leaving it just short of anchor 5's comprehensive coverage.

4 / 5

Distinctiveness Conflict Risk

A single named CLI with load-gating triggers ("Load only when the user explicitly authorizes this skill or names Grok Build / grok-build") gives it a clear niche with minimal overlap risk against any other skill — matches anchor 5, not 4, because no closely related skill would capture these triggers.

5 / 5

Total

18

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

XiaomiMiMo/MiMo-Code

Path

packages/cli/src/skill/builtin/.bundle/grok-build/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.