reddb.io
Official@reddb-io
reddb.io ships RedSkills: a governed agent-memory graph, LLM wiki, and issue-tracker-driven autonomous development command suite for Claude Code and Codex repos.
Agent Skills by reddb.io
Showing 72 vetted skills indexed across 1 GitHub repositories.
reddb-ioinit
Configures per-project memory storage mode and writes the memory plugin config.
reddb-ioimprove-skills
Generate approval-gated skill improvement proposals from skill telemetry failure evidence.
reddb-iohealth
Report Memory plugin operational health via read-only JSON healthcheck diagnostics.
reddb-ioingest
Index repository code and markdown into a memory graph with deterministic extractors.
reddb-iodoctor
Inspect and prune stale nodes in the memory graph after explicit confirmation.
reddb-iowiki
Ingest sources, query, and lint a markdown-based LLM Wiki knowledge base.
reddb-iowiki-init
Bootstraps an LLM Wiki layout with schema, templates, and gitignore rules in a repository.
reddb-iocontext-status
Reports project context stack readiness including agent rules, Memory graph, telemetry, and Wiki status.
reddb-iorecall
Searches project memory notes and returns ranked governed context for a query.
reddb-ioskills-status
Diagnose Skill telemetry status and recent Skill usage events.
reddb-ioview
Opens the project Memory graph store in a visual red-ui workspace or browser Workbench.
reddb-ioextract
Extract durable facts from transcripts into a Memory graph via a configured AI provider.
reddb-ioexport
Export the memory graph to graph.json, interactive graph.html, and audit.md files.
reddb-iostore
Save one durable work fact to the project's configured memory surface.
reddb-iocreate-plugin
Scaffolds new RedSkills repository plugins compliant with marketplace and structural contracts.
reddb-iocapture
Save typed knowledge artifacts into the project Brain store.
reddb-iothink
Synthesize cited answers from project Brain artifacts with confidence levels and evidence gaps.
reddb-iostatus
Reports Brain store initialization state, connection string, and artifact counts.
reddb-iosearch
Search Brain knowledge artifacts for captured decisions, people, plans, and sources.
reddb-iosetup-pre-commit
Configures Husky pre-commit hooks with lint-staged, Prettier, typecheck, and tests.
reddb-iogit-guardrails-claude-code
Configure Claude Code PreToolUse hooks to block destructive git commands before execution.
reddb-iobranch-lock
Locks the agent to a git branch and blocks branch-switching commands via pre-tool hooks.
reddb-iomigrate-to-shoehorn
Migrate TypeScript test files from `as` assertions to @total-typescript/shoehorn helpers.
reddb-iowriting-fragments
Captures unstructured writing fragments into a single markdown document through an interview session.
Frequently Asked Questions About reddb.io
FAQPage SchemaWhat tasks can I automate using reddb.io's RedSkills?▼
RedSkills automates repo memory ingestion and recall, spec-to-ticket breakdown, issue triage, autonomous AFK queue draining with PR landing, TDD red-green loops, merge-conflict resolution, code review against standards and spec, and CDP-driven ground-truth verification of changes.
Who is RedSkills designed for?▼
RedSkills targets software engineers and maintainers running Claude Code or Codex agents in real repositories. It serves operators managing autonomous worker fleets, maintainers curating ADRs and skills, and developers wanting governed long-term project memory across sessions.
How do I install and activate RedSkills in a repository?▼
Run the red-setup skill, the only authorized creator of a repo's .red/ directory. It prompts which plugins to enable (dev, memory, brain), writes activation flags to .red/config.yaml, and sets up AGENTS.md/CLAUDE.md blocks. Hooks stay inert without explicit per-plugin opt-in.
What are the prerequisites for the memory graph features?▼
Memory graph features require running memory init in graph mode first, which provisions a per-project RedDB store. Fact extraction additionally requires a configured AI provider in the plugins.memory block of .red/config.yaml. Markdown-only mode works with no engine.
How does RedSkills keep autonomous agents safe?▼
Safety comes from approval-gated skill improvements, git guardrail hooks blocking destructive commands like push and reset --hard, branch locks pinning agents to one branch, human-in-the-loop issue gates, and ground-truth verification requiring fresh browser snapshots before claiming success.