sefi-core

Plugin sefi-core from marketplace sefi-agents in xsefirosus/sefi-agents. Commands read from the repo's manifest — 26 days ago.

Skills (20)

• anti-hallucination

  Use when producing any factual output -- paths, APIs, numbers, citations, config keys, or claims about code behavior. The canonical no-invention rule every agent points to; unknown lookups become UNKNOWN, unrun executions become PENDING, and every claim traces to a source.

• backend-design

  Use when designing or building anything below the API seam -- endpoints, data models, migrations, background jobs, or service boundaries. Contract-first API design with trust-boundary validation, idempotent mutations, reversible migrations, and an explicit error taxonomy.

• design-style-profiles

  Use when a selected UI direction needs a constrained visual profile to tune tokens, density, depth, borders, typography treatment, or composition.

• expo-native-design

  Use when designing or auditing an Expo or React Native interface after a UI direction is selected.

• focus

  Use when focus_mode.enabled is true in sefi.config.yml. Shapes orchestrator replies to the human for actionability -- lead with the next action, number multi-step work, restate progress every turn, cap lists at five, matter-of-fact errors, no preamble or closers. Governs structure, not phrasing -- p

• frontend-design

  Use when designing, building, auditing, or redesigning any user interface, or studying a design reference. The anti-AI-slop craft skill -- one committed aesthetic direction from a named lane catalog, typography-first hierarchy, a spacing system, tokens, designed states, and non-negotiable accessibil

• loop-engineering

  Use when designing, running, or resuming a loop, or mapping the five moves to repo mechanics. Covers discovery, handoff via worktrees, verification, persistence, and scheduling, plus grep-countable stop conditions, deterministic tripwires, and the git-trust and cycle-count resume rules.

• memory-protocol

  Use when nominating, closing, recovering, searching, or indexing Sefi's local-first Memory Journalist notes.

• motion-design

  Use when a selected UI direction has nontrivial animation, transition interruption, focus behavior, or reduced-motion requirements.

• n8n-workflow-design

  Use when designing an n8n or similar automation workflow as an implementable spec. Opens with the must-haves Rule block covering triggers, idempotency, retries, secrets, webhook security, observability, and cost-per-run, and points to references/node-patterns.md for deep patterns.

• premortem

  Use when a plan needs to survive contact with reality before execution starts, especially high-stakes or high-budget work. Forensic pre-execution failure analysis -- treats the plan as already failed, ranks the 7 likeliest causes of death, delivers a verdict, rebuilds the plan, plays the adversary,

• release-tracking

  Use for every version bump, release, tag, or investigation of a mismatched sefi-agents version. Reconciles six independent publication surfaces -- plugin.json, marketplace.json, CHANGELOG, git tag, GitHub release, GitHub marketplace index -- against one append-only evidence ledger, and reports parti

• retro-improve

  Use when running the self-improvement retro over qa-engineer rejects, gate failures, and metrics to propose bounded edits. Single-writer, append-only-safe self-improvement that edits only managed-by sefi-agents files and stops when improvement is disabled.

• run-sefi-benchmark

  Operate this repository's real-token-spend blinded paired A/B benchmark -- the full sefi-agents chain versus one strong model on frozen cases, needing explicit per-invocation human authorization of its dollar spend before any model call. Use only when a human explicitly asks to run the sefi benchmar

• security-review

  Use when reviewing a diff, dependency change, or configuration that touches a trust boundary before it can reach a PR. The security gate covering secrets, injection surfaces, unsafe constructs, dependency risk, authorization, and data handling.

• sefi-orchestration

  Use when routing a request to the right agent, handing off between agents, or dispatching a subagent. The routing brain covering roster, handoff rules, the parse ladder for structured output, model routing, and pointers to the harness map and never-auto-merge rule. Session-loaded on Claude Code, Ope

• swiftui-design

  Use when designing or auditing a SwiftUI or native Apple interface after a UI direction is selected.

• systems-audit

  Run an on-demand evidence-based audit of Sefi department outputs through Systems Auditor, with bounded scope, honest coverage, capped findings, and an ignored-local report.

• technical-writing

  Use when writing or revising user-facing prose -- READMEs, changelogs, guides, adapter docs, or release notes. Audience-first structure with a quickstart before theory, every claim verified against the repo, and no invented benchmarks or features.

• terse-mode

  Use when terse_mode.enabled is true in sefi.config.yml and a chat reply exceeds roughly 1.5-2k output tokens. Narration compression with an explicit drop-list and never-touch-list; it compresses phrasing, never scope, and never touches code, paths, or safety warnings.

Slash commands (13)

• /sefi-core:audit

  Audit department outputs for one scope through systems-auditor, writing one fenced report without modifying source.

• /sefi-core:close-session

  Close the current substantial work session into one private, structured Memory Journalist note.

• /sefi-core:cross-memory

  Control the optional local-only cross-project memory mirror.

• /sefi-core:init

  Initialize Sefi in the current project root with private local memory, state, loops, and configuration.

• /sefi-core:loop-new

  Interview briefly, then generate a new loops/<name>.loop.md with all six elements filled.

• /sefi-core:map-codebase

  Produce a bounded, evidence-backed local codebase map, trace, impact report, delta, visualization, or context packet.

• /sefi-core:memory-index

  Rebuild or check the disposable local Memory Journalist search index.

• /sefi-core:memory-search

  Search private local Memory Journalist notes, optionally in one explicitly named project.

• /sefi-core:retro

  Run the retro-improve self-improvement pass once, honoring sefi.config.yml.

• /sefi-core:route

  Deterministically load the sefi-orchestration skill and route one task -- for when the auto-trigger might not fire on its own.

• /sefi-core:scout

  Examine an external repository for evidence-backed adoption candidates without copying or implementing it.

• /sefi-core:status

  Print open findings, pending human reviews, and budget headroom (read-only).

• /sefi-core:triage

  Run one discovery turn and write findings to state/triage.md without implementing.

Subagents (17)

• devops-engineer

  Use when the work is pipeline or release mechanics -- CI workflows, worktree lifecycle, scheduled loop wiring, budget enforcement plumbing, or release preparation. Owns the rails the loops run on, and never merges or deploys on its own authority.

• memory-journalist

  Use when a substantial work session needs a private, local-first structured note, recovery, indexing, or an explicit cross-project lookup. Owns the session journal and never writes a raw conversation, secret, command dump, or full diff.

• motion-designer

  Use when a selected interface direction needs nontrivial motion planning, motion reduction, audit, or diagnosis. Owns temporal behavior without changing visual direction or target application code.

• product-manager

  Use when a goal must become an executable spec before any code is written. Turns a goal into a single checkable plan file with a fixed heading skeleton and grep-countable steps, and never implements.

• prompt-engineer

  Use as Stage 0 on an interactive human message, before the engineering-manager opens the routing table. Restates a raw message into unambiguous single-intent statements with only its stated constraints, and never routes, plans, or writes a file.

• qa-engineer

  Use when a built plan slice must be judged before it can be trusted or merged. The adversarial reviewer runs a task-scoped gate against executed evidence and returns PASS or REJECT, never praise.

• research-adoption-scout

  Use only when an explicit adoption, reuse, or third-party evaluation is requested after a codebase map. Produces source, license, and provenance evidence with Adopt, Defer, or Reject; never changes a target, plan, dependency, or third-party material.

• research-analyst

  Use when a task needs external or repository context gathered before planning or implementation. Gathers web, repo, and doc context inside its own window and returns only a bounded digest, never editing files.

• research-codebase-cartographer

  Use only when an explicit codebase map, trace, impact, delta, visual, or context request is made. Reads target source deterministically and writes only declared map artifacts, never changes the target codebase or a plan.

• security-engineer

  Use when a diff, dependency change, or config touching a trust boundary needs a security review before a PR. Runs the security-review gate against the diff and returns findings with severity, never a rubber stamp.

• sefi-agents

  Sefi-Agents - Use when work must be routed to the agent, sequenced across a handoff chain, or dispatched to a subagent. Routes per the routing table, enforces output contracts and budgets, and never edits files or does the work itself.

• software-engineer

  Use when an approved plan slice needs to be built. The full-stack generator implements exactly one plan slice in an isolated worktree, runs the gate before declaring done, and never judges its own quality beyond "gate passed.

• solutions-architect

  Use when a business process needs an automation designed for n8n, Make, GoHighLevel, RAG, or Vapi. Produces an implementable spec after a locked ROI review with equal-weight alternatives; recommends, never deploys.

• support-engineer

  Use when inbox items, issues, or incoming reports need intake, triage, and routing. Reads each item once, classifies actionability, applies consume-before-act on human decisions, and never implements fixes itself.

• systems-auditor

  Audit outputs and write capped report without dispatching, source edits, or fix planning.

• technical-writer

  Use when user-facing prose needs writing or revising -- READMEs, changelogs, guides, or adapter docs. Writes to the technical-writing skill's rules with every claim verified against the repo, and never documents features that do not exist.

• ui-ux-designer

  Use when a user interface needs a design spec, prototype comparison, audit, redesign, or reference study. Owns visual direction per the frontend-design skill and never writes target application code.

Hooks (3)

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}/scripts/inject-memory.sh"

• SessionStart

  "${CLAUDE_PLUGIN_ROOT}/scripts/inject-orchestrator-role.sh"

• PreToolUse

  "${CLAUDE_PLUGIN_ROOT}/scripts/check-bash-write.sh"

Similar plugins