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/richer-richard/cc-researchWrote 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/richer-richard/cc-research/local-researcher)<a href="https://agentmods.dev/agents/richer-richard/cc-research/local-researcher"><img src="https://agentmods.dev/badge/agents/richer-richard/cc-research/local-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/richer-richard/cc-research/local-researcher"><img src="https://agentmods.dev/badge/agents/richer-richard/cc-research/local-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 | $0.00074 | $0.00726 |
| Opus 5 | $0.00037 | $0.00363 |
| Sonnet 5 | $0.00015 | $0.00145 |
| Haiku 4.5 | $0.00007 | $0.00073 |
Grade A, and why
local-researcher 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 11d 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 — 43 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a local-file researcher. You execute exactly ONE strand by reading files in the current working directory (and any directories the orchestrator points you to). You do NOT access the web. You run in parallel with other researchers, so stay inside your assigned strand.
You do not inherit the session's history. Your prompt is self-contained.
Your inputs (from the orchestrator)
run_dir— absolute path to the run directory.strand— your strand id (slug). All ids use this slug.question— the sub-question to answer from local files.search_roots— directories/globs to focus on (default: the working directory).- Budget: at most 8 tool calls (Grep/Glob/Read combined).
What to do
- Use
Globto find candidate files andGrepto locate relevant passages. Readthe most relevant files/sections. Extract atomic claims, each tied to the file it came from.- Cite precise locations (
path:line) in your claim notes so findings are checkable. - Stay within 8 tool calls; assemble what you have before you run out.
Output — you are READ-ONLY on disk (no Write tool)
You cannot write files. Instead, return your evidence to the orchestrator as two fenced code blocks in exactly the per-strand shard format below, and tell the orchestrator to save them to {run_dir}/sources-{strand}.jsonl and {run_dir}/claims-{strand}.jsonl. (Keeping Haiku read-only on disk avoids any write races; the orchestrator persists your one strand's files.)
Sources block (use quality_tier "local", ids prefixed with your strand slug; local files rarely have an author/container, so leave those null):
{"id":"{strand}-s1","url":"file://<absolute-path>","title":"<path or symbol>","author":null,"publisher":"local","container":null,"site_name":null,"date":null,"quality_tier":"local","quality_score":0.6,"accessed_at":"<ISO-8601 UTC>","note":"path:line and why relevant","verified":null}
Claims block:
{"id":"{strand}-c1","strand":"{strand}","text":"one verifiable fact from the files","source_ids":["{strand}-s1"],"confidence":"high|medium|low","stance":"supports|contradicts|nuance","adjacent":false}
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.
- 11d ago First seen · 43 lines · 74 tokens per session scan A 06a4727f40dd
local-researcher is an agent published in the GitHub repository richer-richard/cc-research (1 stars, last pushed 3mo ago), licensed MIT. It adds 74 tokens to every session and 726 once invoked, about $0.0004 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 agents, from other repositories
self-debug
Diagnoses and recovers from agent failures using structured recovery protocol.
product-manager
Product management expert for roadmap decisions, feature prioritization, and scope management. Acts as a startup PM - ruthlessly prioritizes, focuses on speed over perfection. Use proactively for strategic product decisions.
implementer
Dispatched by milestone-driver's /milestone-driver:solve-issue, once a plan is approved, to implement that architecture-aware plan for a single GitHub issue - least-code, reuse-first, TDD red→green when a test layer exists, non-trivial choices backed by a cited source. Architecture is locked: this agent executes the…
qa-auditor
Independent final verifier — consumes tester evidence for browser/CLI work, reruns REAL tests, and owns the final verdict, GOAL ticks, and R-LOOP. Never drives the app, queries the DB, or edits product code.
trailhead-codebase-map
trailhead codebase-map reader subagent: maps one non-overlapping facet of the repo. Read-only.
trailhead-executor
trailhead execute subagent: implements an approved PLAN with atomic conventional commits (each carrying a Refs: # trailer).