القائمة
Run adversarial-review then review on the same target (dual-lens harden recipe)
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Use when the user says to use Grok, wants a Grok review, wants a second opinion from Grok, or asks to delegate a task to Grok. Also use for getting an independent code review from Grok, having Grok reason about an architecture or debugging question, having Grok implement code (live tree by default, or opt-in isolated worktree), or having Grok independently verify someone else's implementation. Trigger phrases include "use grok", "grok review", "second opinion from grok", and "delegate to grok".
Check Grok readiness and optionally toggle stop gate / run mode / Codex agents scope (Codex agents auto-install on SessionStart)
Have Grok implement code (default: live tree; opt-in isolated worktree). Nothing is committed or pushed
One-call delegate: Grok code then auto-handoff verification (verify-only; mode-aware integrate elsewhere)
ACP multi-turn peer channel (start/prompt/stop). Default peer path; opt out with GROK_DISABLE_ACP=1
Read a verified implementation handoff for a completed Grok code run (runId only; no apply)
| name | dual-lens |
| description | Run adversarial-review then review on the same target (dual-lens harden recipe) |
| argument-hint | [--target <path>] [--task <text> | --task-file <path>] |
| allowed-tools | Bash(node:*), Bash(git:*), AskUserQuestion |
SKILL.md and run.mjs).SKILL_BASE to that path. Do not invent versioned cache paths.SKILL_BASE='<Base directory for this skill - absolute path from Skill tool>'
# Required for completion notifications (plugin/references/execution-context.md):
export GROK_COMPANION_EXECUTION_CONTEXT=foreground # or background
node "$SKILL_BASE/run.mjs" <mode> [args...]
run.mjs finds the plugin install from its own location and runs
scripts/grok-companion.mjs. No CLAUDE_PLUGIN_ROOT / PLUGIN_ROOT required.
If the host already exported CLAUDE_PLUGIN_ROOT or PLUGIN_ROOT, you may call
node "$CLAUDE_PLUGIN_ROOT/scripts/grok-companion.mjs" instead; prefer
"$SKILL_BASE/run.mjs" whenever the Skill tool loaded this skill.
Return companion stdout verbatim. Never put free-text in --task "...";
use --task-file - with a single-quoted heredoc.
Use one completion signal for the recipe: pass --no-notify on the first
(adversarial) run so only the final review may notify when notifications are on.
export GROK_COMPANION_EXECUTION_CONTEXT=foreground
node "$SKILL_BASE/run.mjs" adversarial-review --no-notify [flags from "$ARGUMENTS"] --task-file - <<'GROK_TASK'
<operator focus or paste the task>
GROK_TASK
export GROK_COMPANION_EXECUTION_CONTEXT=foreground
node "$SKILL_BASE/run.mjs" review [same --target] --task-file - <<'GROK_TASK'
Confirm or refute high/critical findings from the adversarial pass. Prefer residual risks.
GROK_TASK
Full recipe: repo docs/dual-lens-harden.md.