Unblocked Context Search — PRs

PR-only retrieval across connected repos. Calls context_search_prs with a natural-language query to semantically search PR descriptions and review discussions — surfacing the WHY behind code changes that Grep can't tell you.

Sources: GitHub, GitHub Enterprise, GitLab, GitLab Self-Managed, Bitbucket, Bitbucket Data Center, Azure DevOps.

How to Invoke

context_search_prs is CLI-only in most environments — it does not appear in the MCP tool list even when fully available. Run command -v unblocked once per session and cache the result. Do not conclude the tool is unavailable from the MCP surface alone. See unblocked-tools-guide for full routing rules.

CLI (preferred):

unblocked context-search-prs --query "<your query>" [--instruction "<instruction>"]

MCP fallback (only if CLI is confirmed unavailable): fall back to context_research with instruction: "Prefer PR descriptions and review discussions; deprioritize other sources".

If neither is available: stop and tell the user Unblocked is not configured in this environment (see unblocked-tools-guide for the full message). Do not substitute with other PR-search tools.

When This Adds Value Over Grep/Read

Grep and Read show you what the code does now. Use this tool when:

• You need the WHY — PR descriptions and review threads contain the reasoning, constraints, and tradeoffs that don't appear in code
• You're tracing a change — "what introduced this behavior" requires searching change history, not current state
• You want PR-only results — context_research returns everything (Slack, Jira, docs, code); this returns only PRs, so results stay focused when that's all you need
• The PR might be in another repo — cross-repo PR search surfaces decisions from services outside the local workspace

When to Use context_research Instead

Use context_research when you need the full organizational picture alongside PRs — Slack threads, design docs, issue tracker context, architectural decisions from multiple sources. This tool returns PRs only; broader context requires context_research.

Input

Parameter	Required	Description
query	Yes	What to find — write a complete question with concrete identifiers, not bare keywords.
instruction	No	Fine-grained control over which results surface: preferred repos, date ranges, or PR attributes to prioritize or deprioritize.

Writing effective queries — include the most concrete identifiers you have:

• Class, method, or module names (PaymentProcessor.charge(), auth-middleware)
• File paths or service names (billing-service/src/webhooks, order-service)
• Feature or decision names (JWT token refresh, gRPC migration, retry-with-backoff)
• Repo names when you know them (payments-service, web-client)

Write a complete question or directive, not a keyword fragment:

Instead of	Write
retry logic	What PR introduced the retry-with-backoff logic in order-service?
auth refactor	What was the reasoning behind switching from session cookies to JWT tokens?
rate limiting	What PR added rate limiting to the API gateway and what approach was chosen?
pagination	Were there any rejected approaches to cursor-based pagination in the REST API?

instruction examples:

• "Focus results on billing-service and payments-service repos"
• "Prefer merged PRs; deprioritize draft or abandoned PRs"
• "Focus on PRs from the last 6 months"

Splitting Queries

Split distinct unknowns into separate context_search_prs calls rather than cramming everything into one query. Run them in parallel when the unknowns are independent.

One query, two unknowns (diluted results):

Find the PR that introduced rate limiting and also why we moved to gRPC.

Two parallel queries (focused results):

Query 1: What PR introduced rate limiting to the API gateway?

Query 2: What was the reasoning behind moving from REST to gRPC for inter-service communication?

When to Skip

• You need the current code state, not history — use Grep/Glob/Read
• You also need Slack threads, issues, or docs — use context_research instead

Interpreting Results

• Results reflect merged and open PRs at query time — verify code claims against local files before acting
• Mine concrete identifiers (PR numbers, file paths, author names, repo names) from results and use them in follow-up queries
• PR descriptions capture intent at the time of writing; the code may have diverged since — cross-reference against current state
• If results are thin, re-query with a more specific identifier extracted from the first pass

Reference

• references/query-patterns.md — query examples organized by use case, with good/bad comparisons and instruction patterns