Repository navigation
Feat/agent skill extend commands api - #574
Merged
Merged
Conversation
NRedisStack had no agent instructions for adding commands. Add .agents/skills/extend-commands-api/SKILL.md, with .claude/skills symlinked to it as in Lettuce, modelled on the Jedis and Lettuce skills of the same name and grounded in this repo's conventions: - the module layout: interfaces with <seealso> docs, sync classes inheriting async, <Module>CommandBuilder returning SerializedCommand with a CommandCategory, and the Literals classes - PublicAPI.Unshipped.txt, and Experiments for preview features - tests: EndpointsFixture with AllEnvironments / StandaloneOnly, CreateKeyName for multi-key commands on cluster, SkipIfRedisTheory gating, and REDIS_VERSION - dotnet format, plus FT.ALIASLIST (redis#518) as the worked example Supervised (default) follows plan mode and approval. Unattended mode, used only on `Mode: unattended` or CLIENT_SKILL_MODE=unattended, takes the HLD and the endpoints from the automation. It runs the new theories against standalone and cluster via REDIS_ENDPOINTS_CONFIG_PATH and reports counts per topology.
The automation's standalone and cluster now require passwords. endpoints.json carries them per entry, EndpointConfig.CreateConnection applies them, and the cluster probe authenticates.
mgravell
approved these changes
Oct 5, 2026
mgravell
left a comment
mgravell
left a comment
Collaborator
There was a problem hiding this comment.
I'll need to check this when the v4 work lands
uglide
added a commit
that referenced
this pull request
Oct 9, 2026
…ll (stacked on #574) (#575) * Add an extend-commands-api agent skill (supervised and unattended) NRedisStack had no agent instructions for adding commands. Add .agents/skills/extend-commands-api/SKILL.md, with .claude/skills symlinked to it as in Lettuce, modelled on the Jedis and Lettuce skills of the same name and grounded in this repo's conventions: - the module layout: interfaces with <seealso> docs, sync classes inheriting async, <Module>CommandBuilder returning SerializedCommand with a CommandCategory, and the Literals classes - PublicAPI.Unshipped.txt, and Experiments for preview features - tests: EndpointsFixture with AllEnvironments / StandaloneOnly, CreateKeyName for multi-key commands on cluster, SkipIfRedisTheory gating, and REDIS_VERSION - dotnet format, plus FT.ALIASLIST (#518) as the worked example Supervised (default) follows plan mode and approval. Unattended mode, used only on `Mode: unattended` or CLIENT_SKILL_MODE=unattended, takes the HLD and the endpoints from the automation. It runs the new theories against standalone and cluster via REDIS_ENDPOINTS_CONFIG_PATH and reports counts per topology. * Describe the password-protected unattended servers The automation's standalone and cluster now require passwords. endpoints.json carries them per entry, EndpointConfig.CreateConnection applies them, and the cluster probe authenticates. * docs(agents): add create-implementation-plan-for-redis-api-change skill * docs(agents): address Codex review on the planning skill - folder-to-class stem mapping for CuckooFilter (Cuckoo*) and CountMinSketch (Cms*) - TimeSeries split test classes need both the sync and the async target - add rows may name files not yet on master; decision_class: none for no-change plans; supervised runs default to ./PLAN.md; agent guidance is procedure, not data
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Note
Low Risk
Documentation-only agent skill and symlink; no production code, auth, or runtime behavior changes.
Overview
Adds an agent skill (
extend-commands-api) that guides humans or automation through adding or extending NRedisStack Redis Stack module commands with evidence-first workflow (HLD, server PR,redis-cli), a change-type decision tree (A–E), and maintainer conventions (builders,CommandCategories,PublicAPI.Unshipped.txt, sync/async + standalone/cluster tests).The skill defines supervised (plan + approval) vs unattended (RedisClientsBot-style parity) modes, including env vars, reporting requirements, and boundaries (no CI/version bumps in unattended runs). A symlink under
.claude/skills/points at the same content for Claude tooling.No NRedisStack library or test code changes—only documentation for agents.
Reviewed by Cursor Bugbot for commit a5ab55c. Bugbot is set up for automated code reviews on this repo. Configure here.