To delegate, prompt and manage subagents. MUST activate to spawn subagent with a quality prompt. MANDATORY unless trivial one-liner.
SkillScore73Maintained
Automatic
◆AI Engineer pack
Last verified 9/3/2026
What it changes
Passed 12 of 17 test cases with the skill, and 7 of 17 without it.
Without skill41%
With skill73%
Checks it passed that the model alone missed
The subagent prompt lists `subagent-directives` as a mandatory skill entry
The orchestrator explicitly applies `assets/o-session-execution-controller.md`
Intermediate artifacts and large I/O reports are stored in a dedicated TEMP folder
and 5 more in the full report
7 fixed9 passed either way6 still failing0 broke
Pass-rate lift
+32 pp
Tokens
+47%
Agent turns
0%
Response time
-32%
Model gemini-3.6-flashMeasured 9/3/2026v316 of 22 cases passed with skill · 17 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 7
Prompt
You are formulating a dispatch prompt for a subagent to write integration tests for our payment gateway module. Which skill entry is mandatory under the `Skills*` field in the prompt block?
What was checked, without and with the skill
Checked
Without skill
With skill
The subagent prompt lists `subagent-directives` as a mandatory skill entry
Missed
Passed
case-08 · gemini-3.6-flash · 9/3/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md80 lines3 files
Prerequisites: USE SKILL hitl, load-project-context
MUST use available subagents.
Manage subagents as senior team lead; own orchestration end-to-end. Subagents = your team: fresh context each run, CAN cheat, CANNOT see user; user CANNOT see orchestrator<->subagent channel → trust-but-verify, assume Murphy's law, poka-yoke the process. Adapt management best practices to the specific request. Tell WHAT + HOW-to-think; encourage thinking over mechanical work; never do subagents' tasks yourself — organize them. APPEND to instructions, never paraphrase/duplicate; ground via refs (files/instructions/phases/steps/skills) + MoSCoW; consult advisor/subagent on high-impact/ambiguous/architectural decisions.
Subagent output = input, not truth: judge/reconfirm/fill gaps; spawn focused follow-ups; merge into one grounded result — never blind-accept/discard.
Proactively use available skills, tools, MCPs — incorporate in plan.
Integrate checklists: generate overall end-to-end checklist (like DoD), actualize in planning/discovery, make final reviewer & validator to ultimately honestly check it.
If supported, identify and use long-running (as dedicated area engineer, for follow ups, maintain context) vs one-shot subagents (throw-away fresh reviews).
<request_sizing>
~1-2 files, one area → SMALL → hold it yourself on todo-task ledger; subagents for fresh-eye review.
~up to 10 files, one area → MEDIUM → keep todo-task ledger; build + manage a subagent team — no longer a worker. MUST APPLY SKILL FILE assets/o-team-manager.md.
10+ files or several areas → LARGE → orchestrator context will be overloaded, so it must delegate everything to subagents, todo tasks must be at a higher level and help orchestrator itself.
If requested directly: use session-level EXECUTION_CONTROLLER (plan ⊃ phases ⊃ steps ⊃ tasks) and APPLY SKILL FILE assets/o-session-execution-controller.md.
Complexity may shift one band; re-size as reality changes — discovery, surprises, clarification, target already done.
</request_sizing>
Dispatch:
Subagent prompt MUST use <subagent_prompt_template> — terse, factual, specific, DRY.
Ambiguous → clarify before dispatch.
Lightweight = small clear self-contained task (build, tests) — few skills, no project context. Full = specialized role / larger task — project context + role skills.
Routing:
Parallelize collision-free; make collision-safe: own fileset, git worktree, return-diff-only.