Asks a council of models about a hard problem, when the model calls its tool or you run /council : Claude models, and Gemini models when gemini-core has a key, answer in parallel, and the model of the session writes one verdict from their answers.
docs-gardener
joeldg/claude-mods
Keeps a project's markdown true to the project: deterministic checks (dead references, stale TODOs, duplicated paragraphs, orphans, dated series, oversized files, finished plans) at write time and from a docs_check tool, and /docs audit, which adds a model edit plan you apply item by item from a pane, or all at once for a scheduled run. Only /docs audit calls a model (one Haiku call per...
review-layers
seanjohnson-dev/review-layers
Layered code review for agent-written changes, the way the Claude Code team lands code: checks run first, a risk tier is set from the change, parallel bug hunters post only findings that verifiers confirm, security findings block, review rounds are capped, repeated lessons go into REVIEW.md, incidents become an eval, and git push / gh pr are gated until /land is green. Token-light: haiku by...
verinoda-live
Verinoda-Labs/verinoda-symbiosis
Verinoda in Claude Code: the Verinoda MCP server's tools listed in front of the model with a one-paragraph call rule, the code index kept fresh while Claude edits, an optional check of each edit for names that do not exist, one-line file cards in place of file contents when the conversation is compacted, an improvement list for you to choose from (/verinoda-improve), and a Turkish pane with an...
parallel-guard
u6vu06Ou/parallel-guard
Stops new subagents when too many are running or the machine is low on free memory. Zero model tokens.