research

Research a beads task before implementation.

Command for Claude Code

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 commands/prebid/salesagent/research
Clone the repo
git clone --depth 1 https://github.com/prebid/salesagent

Made for: Claude Code.

Per session 8 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,241 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 $0.00008 $0.01241
Opus 5 $0.00004 $0.00620
Sonnet 5 $0.00002 $0.00248
Haiku 4.5 $0.00001 $0.00124

Measured yesterday against content hash 34a7f4a47d94, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

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.

.claude/commands/research.md · 112 lines

How it starts

The opening of the file, as written. The whole thing — 112 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Research Task: $ARGUMENTS

Instructions

You are researching beads task $ARGUMENTS before implementation begins.

Step 1: Read the Task

Run bd show $ARGUMENTS to get the full task description, acceptance criteria, and any design notes.

Step 2: Read the AdCP Specification (Spec-First)

If the task touches schemas, data models, targeting, protocol behavior, tool inputs/outputs, or buyer-facing fields — read the spec before exploring the codebase. The spec defines what's correct; the code is our current (possibly wrong) implementation of it.

Source of truth: Local adcp repo at /Users/konst/projects/adcp. Read the relevant schema and doc files directly — do NOT rely on training data or assumptions about the spec. Online repo is adcontextprotocol/adcp.

Key locations:

  • JSON schemas: /Users/konst/projects/adcp/static/schemas/source/
    • Core types: core/targeting.json, core/package.json, core/product.json, core/media-buy.json, core/frequency-cap.json, etc.
    • Enums: enums/ (pricing models, statuses, etc.)
    • Media buy operations: media-buy/ (create/update requests/responses, package-request)
    • Signals: signals/
  • Protocol docs: /Users/konst/projects/adcp/docs/ (protocols, media-buy lifecycle, reference)
  • Python types: The adcp Python library may be ahead of the JSON schema (v3 structured geo types exist in Python but not yet in the JSON schema). When they diverge, note the discrepancy in the research artifact.

What to extract from the spec:

  1. Read the relevant JSON schema(s) — identify which fields are spec-defined, their types, and constraints
  2. Check required vs optional — does the spec mandate fields we might treat as optional (or vice versa)?
  3. Check additionalProperties — does the spec allow extra fields?
  4. Note any spec gaps — if the task requires behavior the spec doesn't cover, document it explicitly. Don't invent spec compliance.

Carry these findings into every subsequent step — codebase exploration, engineering checklist, and architecture decisions should all be evaluated against what the spec says.

Read the full file on GitHub · 112 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. yesterday First seen · 112 lines · 8 tokens per session scan A 34a7f4a47d94

Subscribe to this mod's changes

research is a command published in the GitHub repository prebid/salesagent (36 stars, last pushed yesterday), licensed Apache-2.0. It adds 8 tokens to every session and 1,241 once invoked, about $0.0000 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 2026-09-01.