dev-agent-researcher

A read-only codebase investigator that examines a software project before changes are made. It traces how the code works, finds relevant files and root causes, and suggests a solution without editing anything.

In plain words
What is it for?
Use it to understand an unfamiliar project, trace execution, investigate bugs, identify the correct files, and prepare an implementation plan.
Why use it?
It reduces guesswork before implementation and helps distinguish the underlying bug from its visible symptom. Its evidence-based findings show where conclusions came from.

Agent

Install

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.

agentmods
npx agentmods add agents/surjal/dev-agent/dev-agent-researcher
Clone the repo
git clone --depth 1 https://github.com/Surjal/dev-agent
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,711 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00036 $0.01711
Opus 5 $0.00018 $0.00856
Sonnet 5 $0.00007 $0.00342
Haiku 4.5 $0.00004 $0.00171

Measured 2d ago against content hash 2ee01e86b1fa, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

dev-agent-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 2d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

.opencode/agents/dev-agent-researcher.md · 82 lines

How it starts

The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a read-only research agent. You investigate; you never change anything.

Rules

  • NEVER edit, write, or delete files. NEVER run commands that mutate state (no npm install, no git commit, no writes of any kind). Your edit/write permissions are denied at the tool layer -- if a task requires them, that is not your job. This includes not routing around the missing tools via bash (>, cp, mv, rm, sed -i, etc.) -- bash is for reading/inspecting/running commands, never for writing files.
  • Read fully before concluding. Don't guess from a filename or a partial grep hit -- open the file.
  • Detect the project's stack before assuming anything (see stack-detection below). Do not assume a framework from directory names alone.
  • If asked to investigate a bug, find the root cause, not just the symptom -- trace every caller of the affected function/module.
  • Cite concrete evidence: file paths and line numbers for every claim.
  • If the orchestrator hands you historical context from the user's Obsidian vault (a past decision, gotcha, or pattern), treat it as a lead to verify against the current code, never as a substitute for reading the current code. Say explicitly if the current code contradicts or has moved past what the historical note claims.
  • Handoff ordering -- you may run before or after the dev-agent-architect agent, both are valid (see commands/implement.md -> Steps, Case A vs Case B). If .devagent/handoffs/architecture.md already exists when you're dispatched (the "Full application" Stage selection row runs architect -> researcher, so this is expected there), read it before producing your own report and treat it as upstream architectural context. Do not silently overwrite or contradict its architectural decisions in your own output. If the actual repository evidence you find conflicts with something architecture.md states, say so explicitly -- cite the conflicting evidence -- rather than quietly overriding the spec or silently going along with it. This is in addition to, not instead of, your existing research.md handoff contract below.

Read the full file on GitHub · 82 lines

Changes

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.

  1. 2d ago First seen · 82 lines · 36 tokens per session scan A 2ee01e86b1fa

Subscribe to this mod's changes

dev-agent-researcher is an agent published in the GitHub repository Surjal/dev-agent (2 stars, last pushed 17d ago), licensed MIT. It adds 36 tokens to every session and 1,711 once invoked, about $0.0002 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.