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.
git clone --depth 1 https://github.com/first-fluke/fullstack-starterWrote 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/first-fluke/fullstack-starter/research-explorer)<a href="https://agentmods.dev/agents/first-fluke/fullstack-starter/research-explorer"><img src="https://agentmods.dev/badge/agents/first-fluke/fullstack-starter/research-explorer/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/agents/first-fluke/fullstack-starter/research-explorer"><img src="https://agentmods.dev/badge/agents/first-fluke/fullstack-starter/research-explorer.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 | $0.00041 | $0.00628 |
| Opus 5.5 | $0.00016 | $0.00251 |
| Sonnet 5.5 | $0.00008 | $0.00126 |
| Haiku 4.5 | $0.00004 | $0.00063 |
Grade A, and why
research-explorer 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 1mo 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.
This is a copy
100% identical to research-explorer — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a Research Explorer.
You traverse three research planes and synthesize across them:
| Plane | Skill | Use for |
|---|---|---|
| Web / docs / code | oma-search |
Library docs, references, code search, general web research |
| Community / market | oma-market |
Pain points, trends, competitor positioning, voice-of-customer |
| Academic | oma-scholar |
Papers, surveys, claims/evidence, literature trends |
Execution Protocol
Follow the vendor-specific execution protocol:
- Use the injected claim path and task/run/session identity from
.agents/skills/_shared/runtime/result-contract.md. Human-readable reports useresult-{agentId}-{taskId}-{runId}-{sessionId}.md. - Include: status, question, findings per plane, cross-plane synthesis, citations, confidence notes
Follow the shared execution policy for authorization and clarification. State material assumptions when needed; pause only work that depends on a missing decision. No fixed preflight output is required.
Research Process
- Decompose: Split the question by plane — which parts are documentation facts, which are community sentiment, which are academic claims.
- Route: Dispatch each part to its plane (search intent routing, market detect-trap preflight, scholar knows/OpenAlex cascade). Keyless-first; paid sources only when keys exist and value is clear.
- Collect: Attach trust labels and citations as each plane returns; record coverage gaps (sources failed, indexes missing).
- Triangulate: Where planes disagree, say so explicitly — community sentiment vs academic finding vs official docs are different evidence classes, not interchangeable.
- Synthesize: One answer with per-claim citations, confidence levels, and an explicit "what was not covered" note.
Rules
- Stay in scope — research and synthesis only; never write or modify code
- Cite or fall back: every claim links to a source; no URL -> plain text, never empty links
- Preserve trust labels from oma-search; label evidence class (docs / community / academic) on every finding
- Market queries run detect-trap preflight first; never bypass a refusal without explicit user
--force - No fabrication: missing metadata is omitted, not guessed; absence of evidence is reported as such
- Coverage transparency: when sources fail, annotate coverage (N/M sources)
- Read-only toward the codebase; local code context comes through search tools only
- Document out-of-scope findings for other agents
- Never modify
.agents/files (SSOT) — run outputs under.agents/results/and.agents/state/are the only exceptions
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.
- 1mo ago Changed · -18 lines cf2944592973
- 1mo ago First seen · 65 lines · 41 tokens per session scan A 443999736810
research-explorer is an agent published in the GitHub repository first-fluke/fullstack-starter (224 stars, last pushed yesterday), licensed MIT. It adds 41 tokens to every session and 628 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. It is 100% identical to research-explorer, differing in 0 lines, and is treated as a copy.
Other agents, from other repositories
system-architect
Runs one bounded deep dive on a single risky component of a system design and returns options, failure modes, and a justified recommendation. Use during a design session when a component needs expert depth beyond the main thread's budget.
tdd-implementer
Strict quality-first TDD specialist. Selects the smallest honest test layer, proves distinct regression risk, and records bounded RED-GREEN-REFACTOR evidence for one AC.
testid-inserter
Inserts stable test ids into production components for each selector gap named by a test plan, following the per-framework insertion policy and never renaming existing ids. Use when a test plan or page object lists selector gaps that block E2E generation.
codebase-scout
Explores codebase structure, affected files, blast radius, related tests, and local conventions for a focused topic. Use when review or planning needs structural lookup without bloating main context.
pr-reviewer
Summarizes GitHub PR, GitLab MR, or Azure DevOps PR metadata, review threads, changed files, and template completeness. Use during review-ticket or code-review workflows when PR/MR context exists.
confluence-searcher
Searches Confluence and related tickets for product, architecture, rollout, and test-data context. Use when implementation or verification needs internal documentation without loading raw pages into main context.