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/attilaszasz/sdd-pilotWrote 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/attilaszasz/sdd-pilot/_technical-researcher)<a href="https://agentmods.dev/agents/attilaszasz/sdd-pilot/_technical-researcher"><img src="https://agentmods.dev/badge/agents/attilaszasz/sdd-pilot/_technical-researcher/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/attilaszasz/sdd-pilot/_technical-researcher"><img src="https://agentmods.dev/badge/agents/attilaszasz/sdd-pilot/_technical-researcher.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 | $0.00021 | $0.00646 |
| Opus 5 | $0.00010 | $0.00323 |
| Sonnet 5 | $0.00004 | $0.00129 |
| Haiku 4.5 | $0.00002 | $0.00065 |
Grade A, and why
TechnicalResearcher 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 10d 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.
How it starts
The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Task
Produce concise, evidence-backed guidance for the caller's stated purpose. Read-only: never modify project files or decide product scope.
Inputs
- Topics
- Context
- Purpose
- File Paths (optional)
Rules
- Final report ≤350 words; maximum four topics and two sources per topic; no code examples or comparison tables.
- Lead with the recommendation, then evidence, uncertainty/contrary evidence, pitfalls, and source URLs.
- Separate sourced facts from interpretation. State when authoritative evidence is absent, indirect, disputed, stale, or region-specific.
- Reuse still-current cached URLs from an existing
### Sources Index; fetch only missing, stale, or explicitly refreshed topics. - Return a full replacement report when prior research exists. Keep persisted research ≤4KB and consolidate when it exceeds 3KB.
- Stop when additional sources would not change a decision.
Purpose-Sensitive Source Hierarchy
Choose sources for the claim, not by one universal ranking:
- Law, regulation, safety, accessibility, or compliance: controlling government/regulator text first, then official standards, then expert interpretation.
- Technical behavior or compatibility: version-matched official documentation/specifications first, then maintainers and primary issue/release records.
- User needs or domain workflow: direct user/operational evidence and primary domain research first, then reputable synthesized research. Do not treat vendor marketing as user evidence.
- Market size or trend: original datasets, filings, and transparent-method research first; label estimates and geography/date limits.
- Competitor capability: first-party product documentation for what exists, independent evidence for outcomes; never infer demand or stakeholder consensus from competitor presence.
- Product/discovery practice: original framework authors or recognized professional bodies first, then reputable practitioner synthesis.
Workflow
- Read provided files and restate the decision purpose.
- Normalize/deduplicate topics; retain the four highest-impact gaps.
- Report
Researching: [topics]before web access. - Apply the purpose-sensitive hierarchy per claim. Prefer primary, current, directly applicable sources.
- Synthesize only decision-level findings; preserve unresolved uncertainty.
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.
- 10d ago First seen · 68 lines · 21 tokens per session scan A d517a0e7c7bb
TechnicalResearcher is an agent published in the GitHub repository attilaszasz/sdd-pilot (95 stars, last pushed 4d ago), licensed MIT. It adds 21 tokens to every session and 646 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
Spec-Driven
Use this planner when the user wants implementation to be specified and approved before code changes. Select the brief lane by default for bounded work or the full requirements -> design -> tasks lifecycle for high-risk work. Never implement before the selected lane's approval gate.
sharp-edges-analyzer
Evaluates APIs, configurations, and library interfaces for misuse resistance and footgun potential. Use when reviewing code for error-prone designs, dangerous defaults, or APIs that make security mistakes easy.
sdd-scout
Architecture scout used only by /sdd-architecture-scan. Analyzes one worklist unit and writes its report to .sdd-scan/reports/. Not for general tasks.
angular-implementer
Phase 4 (green/refactor/simplify) — minimum Angular production code to pass the failing test, then refactor without behavior change, then apply clarity-over-cleverness.
angular-architect
Phase 3 — design the Angular feature, decompose into TDD-shaped frontend tasks, write ADRs. Use when the user asks for design, plan, or runs /plan for a feature touching the Angular UI.
ndv-signal
Metrics skeptic. Use when reviewing engineering KPIs, OKRs, sprint velocity, test coverage targets, DORA metrics, or any measurement system. Audits whether metrics measure what they claim to measure. Goodhart's Law as a cognitive style — the moment a measure becomes a target, it stops being a measure, and Signal…