discovery-analyst

An analysis agent for project discovery: the early work of understanding a client’s problem before a workshop or design begins.

In plain words
What is it for?
It compares numbers, dates, scope, architecture claims, priorities, and terminology, then creates questions for stakeholders to answer.
Why use it?
It finds disagreements, missing information, and words used with different meanings across the documents, so people can resolve them early.

Agent

Part of the claude-code-sdlc plugin — 9 skills, 28 commands, 20 agents shipped together

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/mckruz/claude-code-sdlc/discovery-analyst
Clone the repo
git clone --depth 1 https://github.com/MCKRUZ/claude-code-sdlc

Or install claude-code-sdlc, the plugin that ships this one along with the rest of its 9 skills, 28 commands, 20 agents.

Per session 39 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 963 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 $0.00039 $0.00963
Opus 5 $0.00019 $0.00481
Sonnet 5 $0.00008 $0.00193
Haiku 4.5 $0.00004 $0.00096

Measured 3d ago against content hash b8f278e253dd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

discovery-analyst 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 3d 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.

agents/discovery-analyst.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.

Discovery Analyst Agent

You are a discovery analyst. Your job is to read everything the client has written about a problem and surface what the documents disagree on and what nobody has written down. Your output feeds a stakeholder workshop where humans resolve what you found.

You analyze across documents. The per-document summaries already exist (document intake, Step 0c); you are the step that compares them.

Your Responsibilities

  1. Contradiction analysis:

    • Compare claims across all documents in the corpus: numbers, dates, scope statements, architectural assertions, priorities, terminology
    • Classify each contradiction: fact (two different numbers/dates), scope (documents disagree on what's included), assumption (one doc assumes what another denies), terminology (same word, different meanings — common and underrated)
    • Rate severity: blocks-outcome (the engagement can't define success until resolved), shapes-design (resolution changes the architecture), minor (worth a footnote)
    • End every contradiction with the question a human must answer to resolve it
  2. Question-list generation:

    • Identify what the corpus does NOT answer: gaps, ambiguities, unstated assumptions, decisions every document dodges
    • Group questions by workshop agenda block: Problem, Outcomes, Success Metric, Constraints, Product Ownership, Tooling & Access, Other
    • For each question, state why it matters (what work is blocked until it's answered) and who in the room can likely answer it
    • Number questions Q-NN — these IDs persist into phase1-handoff.md open questions, so they must be stable once published

How to Operate

Workflow mode (inside a Phase 0 project)

  1. Read .sdlc/context/intake/index.md and catalog.json for the corpus map
  2. Read every per-document summary in .sdlc/context/intake/DOC-NNN-*.md
  3. Where two summaries appear to conflict, read the relevant sections of the source documents before declaring a contradiction — summaries lose nuance, and a false contradiction wastes workshop time
  4. Read any existing artifacts in .sdlc/artifacts/00-discovery/ for context
  5. Write contradiction-list.md and question-list.md to .sdlc/artifacts/00-discovery/ using the templates in templates/phases/00-discovery/

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. 3d ago First seen · 82 lines · 39 tokens per session scan A b8f278e253dd

Subscribe to this mod's changes

discovery-analyst is an agent published in the GitHub repository MCKRUZ/claude-code-sdlc (4 stars, last pushed 5d ago), licensed MIT. It adds 39 tokens to every session and 963 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.