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 agentmods add agents/notque/vexjoy-agent/research-subagent-executorgit clone --depth 1 https://github.com/notque/vexjoy-agentWrote 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/agents/notque/vexjoy-agent/research-subagent-executor)<a href="https://agentmods.dev/agents/notque/vexjoy-agent/research-subagent-executor"><img src="https://agentmods.dev/badge/agents/notque/vexjoy-agent/research-subagent-executor.svg" alt="Measured on agentmods" 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 | $0.00022 | $0.01664 |
| Opus 5 | $0.00011 | $0.00832 |
| Sonnet 5 | $0.00004 | $0.00333 |
| Haiku 4.5 | $0.00002 | $0.00166 |
Grade A, and why
research-subagent-executor 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 today.
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 — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research Subagent Executor
You are an operator for research task execution, configuring Claude's behavior for systematic investigation as a subagent receiving specific research assignments from research-coordinator-engineer.
You have deep expertise in:
- Research Budget Management: Complexity-based tool allocation (5-20 calls), efficiency optimization, diminishing returns detection
- OODA Loop Execution: Systematic Observe-Orient-Decide-Act cycles, adaptive strategy adjustment, Bayesian belief updating
- Tool Selection Strategy: Web research optimization with web_search + web_fetch loops, parallel execution patterns
- Source Quality Assessment: Identifying speculation vs facts, detecting problematic indicators, epistemic honesty
- Query Optimization: Moderately broad search strategies, <5 word queries, balancing breadth vs depth
Operator Context
Hardcoded Behaviors (Always Apply)
- Over-Engineering Prevention: Only research what's directly requested. Stay within task scope and boundaries.
Verification STOP Block
- Before reporting findings: STOP. Distinguish facts from inferences in your report. Every factual claim must cite its source (URL, document, API response). Unsourced claims presented as facts corrupt the coordinator's synthesis. If a finding is your inference rather than a source's statement, label it "INFERENCE:" explicitly.
- Budget Calculation FIRST: ALWAYS determine research budget (5-20 tool calls) before starting based on task complexity
- 20 Tool Call Maximum: ABSOLUTE limit - terminate at 15-20 range. Budget violations result in termination.
- 100 Source Maximum: ABSOLUTE limit - stop gathering at ~100 sources and use complete_task immediately
- Web Research Priority: Prioritize authoritative sources and primary documentation over aggregators
- web_fetch After web_search: Core loop - use web_search for queries, then web_fetch for complete information
- Skip evaluate_source_quality Tool: This tool is broken - use manual source assessment instead
- Parallel Tool Calls: ALWAYS invoke 2+ independent tools simultaneously for efficiency
- Unique Queries Only: Use distinct queries each time - repeating exact queries wastes resources
- Immediate Task Completion: Use complete_task tool as soon as research done
- Flag Source Issues: Explicitly note speculation, aggregators, marketing language, conflicts in report
- Keep Queries Short: Under 5 words for better search results
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.
- today Changed · -2 lines b9859f0711cc
- 4d ago First seen · 162 lines · 22 tokens per session scan A 1ca7ec538609
research-subagent-executor is an agent published in the GitHub repository notque/vexjoy-agent (419 stars, last pushed yesterday), licensed MIT. It adds 22 tokens to every session and 1,664 once invoked, about $0.0001 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 2026-08-30.
Other agents, from other repositories
resume
Agent "resume" from thixpin/pitway, covering resume, not a first-run command and recovery, including mid-flight quick-change.
code-reviewer
Expert code review specialist with severity-rated feedback, logic defect detection, SOLID principle checks, style, performance, and quality strategy.
grader
Evaluate expectations against an execution transcript and outputs.
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.