Use shell completers to suggest valid values to agents
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 43/100
- Issue type
- Feature
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- typescript
- Domain
- api, cli, developer-experience
Research direction
Start with lib/complete.ts and read the linked cli-first-mcp design and tasks, especially tasks 5.1 and 5.2; inspect the per-command tool validation and existing shell completion tests. Trace how completions use the SQLite cache and live API before adding ranked candidates to validation errors and static choices to generated schemas. Done when the shell completion tests pass unchanged and the specified validation tests cover slugs, issue IDs, dataset fields and aggregates, and query keys and values.
Written by the indexing model from the issue text.
Description
Part of #1357. Plan: cli-first-mcp tasks 5.1 and 5.2 (design, tasks). Depends on #1890.
Why
Agents get slugs, IDs, fields, and query keys wrong. The shell completion system already knows valid values. With no search_sentry tool to host a complete mode, completions reach agents through the per-command tools themselves.
Scope
- Split the completers in
lib/complete.tsinto completers over a data source: the SQLite cache with a live API fallback on stdio, the live API uncached on hosted. - When a per-command tool gets an invalid or unknown value (org/project slug, issue ID, dataset field or aggregate, query key or value), return the error with ranked candidates and their descriptions so the agent can retry.
- Put static choices (enum-like flag values) into the tool schema as
examplesat catalog generation time.
Done when
- Shell completion tests pass unchanged.
- Tests for slugs, issue IDs, dataset fields and aggregates, and query keys and values in validation errors.
via @BYK.
--
- Dominant language
- TypeScript
- Stars
- 961
- Forks
- 179
- Avg merge
- 15h 33m
- Merged PRs (30d)
- 476
Getting set up
This project ships no dev container, Dockerfile or contributing guide, so setting up is up to you: start from its README, and see our first-contribution guide for the general steps.
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from getsentry/toolkit
-
MCP mcp-feedback-pattern
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
getsentry/toolkit#2276 · 1 comment ·
Maintainers usually reply within 1 day
-
CLI enhancement MCP
Difficulty 2/5 1-3 hours Newbie friendliness 63/100
getsentry/toolkit#1908 · 1 comment ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
getsentry/toolkit#1471 · 1 comment ·
Maintainers usually reply within 1 day
-
get_sentry_resource gives no guidance for Explore > Errors list URLsPossibly taken @sentry-junior claimed this today. OpenMCP mcp-feedback-pattern
Difficulty 2/5 1-3 hours Newbie friendliness 32/100
getsentry/toolkit#2278 · 1 comment ·
Maintainers usually reply within 1 day
-
Difficulty 3/5 1-2 days Newbie friendliness 65/100
getsentry/toolkit#2275 · 1 comment ·
Maintainers usually reply within 1 day
All issues in getsentry/toolkit
Similar issues
-
Add: BOM CineOpenapproved check:passed streams:add
Difficulty 1/5 Under an hour Newbie friendliness 72/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Maintainers usually reply within 1 day
-
bug platform-android
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
Termix-SSH/Mobile#146 · 1 comment ·
Maintainers usually reply within 2 days
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 66/100
devswha/herdr-web-ui#3875 ·
-
Difficulty 1/5 Under an hour Newbie friendliness 88/100