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 commands/l3digitalnet/claude-code-plugins/researchgit clone --depth 1 https://github.com/L3DigitalNet/Claude-Code-PluginsWhat 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 | $0.00054 | $0.00974 |
| Opus 5 | $0.00027 | $0.00487 |
| Sonnet 5 | $0.00011 | $0.00195 |
| Haiku 4.5 | $0.00005 | $0.00097 |
Grade A, and why
research 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 yesterday.
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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/qdev:research
Research a topic, task, or technology before designing or building, by dispatching the qdev-researcher subagent.
Why this is a subagent
The research workflow performs 6-8 queries through a Tavily-first recall path, Brave/Serper cross-checks, 3-5 full-page Tavily extracts, and per-library Context7 round-trips. Running it in Opus context burns ~25K tokens per sweep on raw search results alone. The Sonnet subagent consolidates research + corroboration + synthesis into one dispatch and returns a compact structured report. This is the v1.3.0 subagent-extraction pattern: the orchestrator stays out of raw search results and receives only the compact structured report.
How to run it
-
Establish topic.
-
If
$ARGUMENTSis provided, use it as the topic. -
Otherwise, gather context with one bash call:
git log --oneline -5 2>/dev/null || trueRead
CLAUDE.mdat the project root if present. From git history, project files, and conversation context, infer the focus area with reasonable confidence. -
If the topic still cannot be inferred, use
AskUserQuestionwith a single bounded question (no two-step pattern):- header:
"Research topic" - question:
"What should I research? (Pick a recent context or use Other to type a topic.)" - options: up to 3 inferred candidates from git/CLAUDE.md context. The implicit "Other" entry lets the user type a free-text topic.
If no candidates can be inferred at all and the user does not provide one, emit
No topic provided.and stop. - header:
Announce:
Research topic: <topic> -
-
Dispatch
qdev-researcherwith the topic.Use the
Agenttool withsubagent_type: qdev:qdev-researcherand a prompt like:Research
<topic>. Default depth=standard. The research-KB scripts live in${CLAUDE_PLUGIN_ROOT}/scripts/; pass that absolute path to the agent asSCRIPTSso it can invokeuv run "$SCRIPTS/build_research_index.py","$SCRIPTS/validate_research_frontmatter.py", and"$SCRIPTS/dedup.py". Run the Tavily-first search path, route library docs through the Context7 gate, corroborate footguns across 2+ sources, run at most one follow-up pass for thin angles, run the reporting cycle (preflight index → dedup → write report with frontmatter → self-validate → regenerate index), and return the structured report per your output format.
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.
- yesterday First seen · 66 lines · 54 tokens per session scan A e1cf86ce2b74
research is a command published in the GitHub repository L3DigitalNet/Claude-Code-Plugins (6 stars, last pushed 2d ago), licensed MIT. It adds 54 tokens to every session and 974 once invoked, about $0.0003 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 2026-08-31.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.