cue-ai 0.5.0 → 0.7.0

package/profiles/ecc/profile.yaml

@@ -0,0 +1,39 @@
+name: ecc
+icon: "🦅"
+description: Everything Claude Code patterns — code review, dev workflow, language-specific rules
+persona: |
+  You're a senior code reviewer who has seen every common failure mode in
+  TypeScript, Python, and Bash. Your job is to catch what the author missed,
+  not to praise the parts that work.
+
+  - **Default to security review.** Every diff gets scanned for injection,
+    secrets-in-logs, missing input validation, unsafe shell quoting.
+  - **Spec-compliance over taste.** Anthropic's SKILL.md spec is the source
+    of truth, not your stylistic preferences. Reference the spec when you
+    push back.
+  - **Suggest the smallest fix that works.** Avoid "you should refactor this
+    whole module" feedback when a 3-line change is enough.
+  - **Format: rule citation → one-line explanation → exact replacement.**
+    Maintainers should be able to apply your suggestion without thinking.
+inherits: core
+skills:
+  local: []
+rules:
+  - common/code-review
+  - common/coding-style
+  - common/development-workflow
+  - common/git-workflow
+  - common/security
+  - common/testing
+  - typescript/coding-style
+  - typescript/patterns
+  - typescript/security
+  - typescript/testing
+commands:
+  - code-review
+  - checkpoint
+  - feature-dev
+  - build-fix
+  - cost-report
+  - aside
+# hooks: inherited from core — bash-quality-preflight + session-summary

package/profiles/event-design/profile.yaml

@@ -0,0 +1,10 @@
+name: event-design
+icon: "💌"
+description: Event collateral design — wedding invitations, save-the-dates, programs, menus, place cards. Bespoke HTML → print-ready PNG, multilingual, fully local
+inherits: core
+skills:
+  local:
+    - event-design/wedding-invitations    # pointer to wyx-sg/wedding-invitation-skill
+    - design/high-end-visual-design       # generic aesthetic / composition guidance
+    - design/brandkit                     # palette / typography decisions
+mcps: []

package/profiles/fleet-control/README.md

@@ -26,7 +26,7 @@ Fleet Control is for coordinated agent runs, Colony state, OMX workflows, guarde
 
 - `claude-mem`
 - `colony`
-- `soul-skills`
+- `cue-skills`
 
 ## Sample Tasks
 

package/profiles/frontend/profile.yaml

@@ -2,6 +2,20 @@ name: frontend
 icon: "🦋"
 description: Frontend UI implementation, redesign, screenshots, testing, and visual polish
 inherits: core
+persona: |
+  You're a senior frontend engineer with a designer's eye. You write React /
+  Next.js / Svelte without copy-pasting from old projects; every component
+  ships with explicit prop types, no `any`, no unhandled loading/error states.
+
+  - **Verify visually, not just structurally.** A passing build doesn't mean
+    the UI works. Open the page in a browser (or screenshot via the screenshot
+    skill) and confirm the change renders as intended.
+  - **Accessibility is a correctness bar.** Every interactive element gets a
+    proper role, label, and keyboard path. WCAG AA is the minimum.
+  - **Hydration mismatch is the silent killer.** Anything that depends on
+    `window`, `Date.now()`, or `Math.random()` at render time is suspect.
+  - **CSS scope or pay the tax.** Default to CSS Modules / Tailwind utility
+    classes; global styles need an explicit reason.
 skills:
   local:
     - design/design-taste-frontend

package/profiles/full/README.md

@@ -22,7 +22,7 @@ Full is a diagnostic / fallback only profile. It is intentionally expensive beca
 - `polymarket-live`
 - `recodee`
 - `ruflo`
-- `soul-skills`
+- `cue-skills`
 
 ## Sample Tasks
 

package/profiles/go-api/profile.yaml

@@ -5,4 +5,5 @@ inherits: core
 skills:
   local:
     - review/code-review
+    - rust/chisel-tool           # Chisel — agent precision file edits (recipe + pointer guide)
 mcps: []

package/profiles/marketing/profile.yaml

@@ -1,6 +1,6 @@
 name: marketing
 icon: "🦜"
-description: Marketing — copywriting, SEO, CRO, growth, channels, brand, awesome-list outreach, X/Twitter
+description: Marketing — copywriting, SEO, CRO, growth, channels, brand, awesome-list outreach, X/Twitter, Reddit, paid ads audit, AI video
 inherits: core
 skills:
   local:
@@ -9,7 +9,18 @@ skills:
     - design/high-end-visual-design
     - design/imagegen-frontend-web
     - meta/awesome-list-submit
+    - higgsfield/higgsfield-generate
+    - higgsfield/higgsfield-product-photoshoot
+    - higgsfield/higgsfield-marketplace-cards
+  npx:
+    - repo: AgriciDaniel/claude-ads
+      skills: [claude-ads]
 mcps:
   - x-api
+  - reddit
+  - google-ads-mcp
+  - meta-ads
+  - mapi-devdocs
+  - Higgsfield
 plugins:
   - marketing-skills@claude-code-skills

package/profiles/predict-everything/profile.yaml

@@ -0,0 +1,9 @@
+name: predict-everything
+icon: "🔮"
+description: Multi-agent prediction & simulation — digital sandbox rehearsals, swarm intelligence, "what if" scenarios via MiroFish / OASIS-style engines
+inherits: core
+skills:
+  local:
+    - predict-everything/mirofish     # 666ghj/MiroFish — prediction engine on top of OASIS
+    - research/awesome-rust-search    # general discovery (useful for finding adjacent sim libs)
+mcps: []

package/profiles/rust/profile.yaml

@@ -1,8 +1,27 @@
 name: rust
 icon: "🦀"
-description: Rust development — cargo, async, traits, error handling, CLI tools, systems programming
-inherits: core
+description: General Rust development — async, serde, tracing, perf, docs, testing, releases
+inherits: rust-core
 skills:
   local:
-    - review/code-review
+    # Runtime + serialization stack (everything non-trivial pulls these in)
+    - rust/async-tokio
+    - rust/serde
+    - rust/tracing
+    # Docs
+    - rust/mdbook
+    - rust/cargo-readme
+    # Perf
+    - rust/cargo-flamegraph
+    # Quality (beyond the gates in rust-core)
+    - rust/cargo-hack
+    - rust/cargo-mutants
+    - rust/cargo-fuzz
+    - rust/snapshot-testing
+    - rust/property-testing
+    - rust/typos-spellcheck
+    # Releases + version mgmt
+    - rust/release-plz
+    - rust/cargo-msrv
+    - rust/cross-compile
 mcps: []

package/profiles/rust-cli/profile.yaml

@@ -0,0 +1,14 @@
+name: rust-cli
+icon: "⌨️"
+description: Rust command-line tools — clap, ratatui TUIs, snapshot testing, cross-platform releases
+inherits: rust-core
+skills:
+  local:
+    - rust/clap-cli
+    - rust/ratatui-tui
+    - rust/snapshot-testing      # the killer feature for CLI output tests
+    - rust/cross-compile         # ship binaries for every target
+    - rust/serde
+    - rust/release-plz           # tagged releases + crates.io
+    - rust/typos-spellcheck
+mcps: []

package/profiles/rust-core/profile.yaml

@@ -0,0 +1,35 @@
+name: rust-core
+icon: "🦀"
+description: Foundation shared by every rust sub-profile — toolchain essentials, gates, discovery
+inherits: core
+skills:
+  local:
+    - review/code-review
+    - research/awesome-rust-search   # crate discovery against awesome-rust catalog
+    # Toolchain essentials — everyone needs these
+    - rust/cargo-basics
+    - rust/clippy-and-fmt
+    - rust/cargo-nextest
+    - rust/bacon-watch
+    - rust/cargo-edit
+    - rust/cargo-expand
+    - rust/sccache
+    # Always-on safety + ergonomics
+    - rust/cargo-audit
+    - rust/error-handling
+    - rust/just-runner
+    - rust/chisel-tool           # Chisel — agent precision file edits (recipe + pointer guide)
+rules:
+  - rust/coding-style
+  - rust/patterns
+  - rust/security
+  - rust/testing
+  - rust/hooks
+commands:
+  - rust-preflight     # /rust-preflight — fmt + clippy + test + audit aggregator
+  - rust-build         # /rust-build — fix build errors incrementally
+  - rust-review        # /rust-review — deep code review
+  - rust-test          # /rust-test — TDD enforcement
+hooks:
+  - cargo-publish-guard.json   # block `cargo publish` without explicit user OK
+mcps: []

package/profiles/rust-embedded/profile.yaml

@@ -0,0 +1,11 @@
+name: rust-embedded
+icon: "📟"
+description: Rust on microcontrollers — probe-rs, embassy async, no_std patterns, cross-compile to ARM/RISC-V
+inherits: rust-core
+skills:
+  local:
+    - rust/embedded              # probe-rs + cargo-embed + defmt
+    - rust/no-std                # std-free patterns
+    - rust/async-tokio           # transferable concepts for embassy
+    - rust/cross-compile
+mcps: []

package/profiles/rust-ffi/profile.yaml

@@ -0,0 +1,13 @@
+name: rust-ffi
+icon: "🔗"
+description: Rust foreign-function interface — Python (pyo3), Node (napi-rs), Kotlin/Swift (uniffi), C (bindgen/cbindgen)
+inherits: rust-core
+skills:
+  local:
+    - rust/pyo3
+    - rust/napi-rs
+    - rust/uniffi
+    - rust/bindgen               # C → Rust
+    - rust/cbindgen              # Rust → C
+    - rust/cross-compile         # FFI almost always means multi-platform
+mcps: []

package/profiles/rust-game/profile.yaml

@@ -0,0 +1,11 @@
+name: rust-game
+icon: "🎮"
+description: Rust game / simulation development — Bevy ECS, perf profiling, asset pipelines
+inherits: rust-core
+skills:
+  local:
+    - rust/bevy
+    - rust/cargo-flamegraph      # perf is the whole game
+    - rust/cross-compile         # ship for win/mac/linux + wasm
+    - rust/wasm-rust             # browser builds via wasm-bindgen
+mcps: []

package/profiles/rust-wasm/profile.yaml

@@ -0,0 +1,11 @@
+name: rust-wasm
+icon: "🕸️"
+description: Rust → WebAssembly — wasm-pack libraries, Trunk/Dioxus SPAs, no_std considerations
+inherits: rust-core
+skills:
+  local:
+    - rust/wasm-rust             # wasm-pack + trunk + dioxus
+    - rust/no-std                # wasm32-unknown-unknown is effectively std-free
+    - rust/serde
+    - rust/typos-spellcheck
+mcps: []

package/profiles/rust-web/profile.yaml

@@ -0,0 +1,17 @@
+name: rust-web
+icon: "🕸️"
+description: Rust web services — axum HTTP APIs, reqwest, sqlx, serde, tracing, container deploys
+inherits: rust-core
+skills:
+  local:
+    - rust/axum-api
+    - rust/reqwest
+    - rust/serde
+    - rust/tracing
+    - rust/sqlx-cli
+    - rust/snapshot-testing      # API response snapshots
+    - rust/release-plz
+    - rust/cargo-chef            # Docker layer caching
+    - rust/typos-spellcheck
+    - rust/async-tokio
+mcps: []

package/profiles/schema.json

@@ -1,6 +1,6 @@
 {
   "$schema": "http://json-schema.org/draft-07/schema#",
-  "$id": "https://github.com/recodeee/soul/profiles/schema.json",
+  "$id": "https://github.com/opencue/cue/profiles/schema.json",
   "title": "Soul Profile",
   "description": "Schema for profiles/<name>/profile.yaml — bundles skills, MCPs, and env for a Claude Code or Codex session.",
   "type": "object",
@@ -16,7 +16,7 @@
       "type": "string",
       "minLength": 1,
       "maxLength": 200,
-      "description": "One-line description shown in `soul list` and embedded in the materialized CLAUDE.md stamp."
+      "description": "One-line description shown in `cue list` and embedded in the materialized CLAUDE.md stamp."
     },
     "icon": {
       "type": "string",
@@ -50,7 +50,7 @@
               { "$ref": "#/$defs/AgentScopedRef" }
             ]
           },
-          "description": "Paths relative to soul/skills/ (e.g. 'medusa/building-with-medusa')."
+          "description": "Paths relative to cue/skills/ (e.g. 'medusa/building-with-medusa')."
         },
         "npx": {
           "type": "array",
@@ -67,7 +67,7 @@
         ]
       },
       "default": [],
-      "description": "MCP server IDs that must exist in soul/mcps/configs/. Accepts plain string or {id, agents} object."
+      "description": "MCP server IDs that must exist in cue/mcps/configs/. Accepts plain string or {id, agents} object."
     },
     "plugins": {
       "type": "array",
@@ -87,6 +87,46 @@
       "type": "object",
       "additionalProperties": { "type": "string" },
       "description": "Env vars set when this profile is materialized; values may contain ${VAR} placeholders."
+    },
+    "rules": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 },
+      "default": [],
+      "description": "Rule files (relative to resources/rules/ or absolute) loaded into the session. Enforce coding standards, security policies, etc."
+    },
+    "commands": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 },
+      "default": [],
+      "description": "Slash command files (relative to resources/commands/) scoped to this profile."
+    },
+    "hooks": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 },
+      "default": [],
+      "description": "Hook scripts (relative to resources/hooks/) activated for this profile. Lifecycle: PreToolUse, PostToolUse, SessionStart, SessionEnd, Stop."
+    },
+    "persona": {
+      "type": "string",
+      "description": "Phase 1: multi-line role-priming text injected at the top of CLAUDE.md. Defines who the agent IS (a senior Rust engineer who defaults to safety), not just what tools it has."
+    },
+    "playbooks": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 },
+      "default": [],
+      "description": "Phase 2: markdown playbooks (relative to resources/playbooks/) — proven step-by-step protocols for common tasks (ship-feature, triage-bug). Symlinked into runtime, indexed in CLAUDE.md."
+    },
+    "qualityGates": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 },
+      "default": [],
+      "description": "Phase 3: Stop-hook validator scripts (relative to resources/quality-gates/) that veto 'done' if work doesn't meet the profile's bar (tests pass, lint clean, etc.)."
+    },
+    "evals": {
+      "type": "array",
+      "items": { "type": "string", "minLength": 1 },
+      "default": [],
+      "description": "Phase 4: eval scenario refs (relative to resources/evals/) declaring 'for task X this profile should be able to handle it'. `cue eval-behavior` runs structural checks."
     }
   },
   "$defs": {

package/profiles/trendradar/profile.yaml

@@ -0,0 +1,11 @@
+name: trendradar
+icon: "📡"
+description: TrendRadar — news aggregation, trend analysis, AI-powered filtering, and notification dispatch via MCP
+inherits: core
+skills:
+  local:
+    - research/trendradar
+    - research/keyword-research
+mcps:
+  - trendradar
+plugins: []

package/resources/mcps/README.md

@@ -1,181 +1,56 @@
-```
-   ┌───────────────────────────────────────────────────────────────┐
-   │                                                               │
-   │     r e c o d e e e  /  m c p s                               │
-   │     ───────────────────────────                               │
-   │     13 MCP servers · linked skills · sanitized configs        │
-   │                                                               │
-   └───────────────────────────────────────────────────────────────┘
-```
-
-> **My laptop's MCP registry.** This repo tracks the MCP servers used by
-> Codex, Claude, and VS Code without committing raw settings or credentials.
-> It also maps each MCP to the skills that depend on it.
-
----
-
-## Why this exists
-
-MCP config drifts even faster than skills: one agent has a server enabled,
-another points at a local binary, VS Code has a separate entry, and the source
-repo link gets forgotten. This repo is the audit layer.
-
-It records:
-
-- which MCP servers are configured
-- where they come from
-- which commands or hosted endpoints launch them
-- which environment variable names they expect
-- which agent skills are related to each server
-
-It does **not** try to be the live secret store. Tokens stay in local config,
-environment variables, password managers, or service-specific credential files.
-
----
-
-## Layout
-
-```text
-configs/
-  codex.sanitized.json      # ~/.codex/config.toml MCP block, redacted
-  claude.sanitized.json     # ~/.claude/settings.json MCP block, redacted
-  vscode.sanitized.json     # VS Code mcp.json, redacted
-  mcp-skill-rules.json      # explicit source links and skill matching rules
-  mcp-skill-map.json        # generated machine-readable MCP -> skills map
-docs/
-  health.md                 # generated local health summary
-  inventory.md              # generated human inventory
-  mcp-skill-map.md          # generated human MCP -> skills map
-mcps/
-  <mcp-name>/README.md      # per-MCP install and health notes
-  <mcp-name>/skills.md      # per-MCP related skills and source link
-  omx/                      # OMX MCP family grouped under one tree
-scripts/
-  check-no-secrets.sh       # fail if tracked docs/configs leak secret-shaped data
-  check-mcp-health.py       # generate docs/health.md without network checks
-  refresh-all.sh            # snapshot configs, map skills, health, scan, diff
-  install-codex-mcps.py     # install placeholder MCP blocks into Codex config
-  install-claude-mcps.py    # install placeholder MCP blocks into Claude settings
-  snapshot-mcps.py          # refresh sanitized MCP config inventory
-  snapshot-mcp-skills.py    # refresh MCP -> skills pages from recodeee/skills
-templates/
-  codex.mcp.toml            # placeholder-only Codex MCP config block
-  claude.mcp.json           # placeholder-only Claude MCP config block
-```
-
----
-
-## MCP Inventory
-
-| MCP | Source | Runtime surface |
-| --- | --- | --- |
-| `Higgsfield` | `https://mcp.higgsfield.ai/mcp` | hosted MCP endpoint |
-| `MedusaDocs` | `https://docs.medusajs.com/mcp` | hosted docs MCP endpoint |
-| `colony` | `https://github.com/recodeee/colony` | local CLI MCP |
-| `coolify` | `https://github.com/StuMason/coolify-mcp` | local npm MCP |
-| `gitea` | `https://github.com/seepine/gitea-mcp` | local binary MCP |
-| `hostinger-api` | `https://github.com/hostinger/api-mcp-server` | local npm MCP |
-| `letsfg` | `https://github.com/LetsFG/LetsFG` | local npm MCP |
-| `react-grab-mcp` | `https://www.npmjs.com/package/@react-grab/mcp` | npm MCP |
-| `omx/*` | `https://github.com/NagyVikt/oh-my-codex` | local Node MCP family |
-
-OMX servers are grouped under:
-
-```text
-mcps/omx/
-  code-intel/
-  memory/
-  state/
-  trace/
-  wiki/
-```
-
----
+# cue/mcps — MCP Server Registry
 
-## Refresh
+> The MCP server configurations used by [cue](https://github.com/opencue/cue). Each profile references servers from this registry by ID.
 
-Refresh sanitized MCP config snapshots:
+## What's here
 
-```sh
-./scripts/snapshot-mcps.py
 ```
-
-Refresh MCP-to-skill mapping from the sibling `recodeee/skills` checkout:
-
-```sh
-./scripts/snapshot-mcp-skills.py
-```
-
-Common full refresh:
-
-```sh
-./scripts/refresh-all.sh
+mcps/
+├── configs/              Sanitized JSON configs (no secrets)
+│   ├── claude.sanitized.json
+│   ├── claude_runtime.sanitized.json
+│   └── codex.sanitized.json
+├── mcps/                 Per-server documentation
+│   ├── gbrain/           Personal wiki with embeddings
+│   ├── colony/           Multi-agent coordination
+│   ├── coolify/          Deployment management
+│   ├── higgsfield/       AI image/video generation
+│   └── ...               13+ servers
+├── plugins/              Claude Code plugin configs
+├── scripts/              Sync and maintenance scripts
+└── templates/            New MCP server templates
 ```
 
-Review diffs before pushing. If a secret appears, remove it before commit and
-tighten the redaction script.
+## How cue uses this
 
-Install placeholder MCP blocks only after reviewing the templates:
+When you run `cue use backend`, cue reads `configs/claude.sanitized.json` to find the MCP server entries for that profile's declared MCPs. The materializer writes them into the runtime's `settings.json`.
 
-```sh
-./scripts/install-codex-mcps.py --dry-run
-./scripts/install-claude-mcps.py --dry-run
+```yaml
+# profiles/backend/profile.yaml
+mcps:
+  - coolify        # ← looked up in configs/claude.sanitized.json
+  - gbrain         # ← inherited from core
 ```
 
-The install scripts write placeholders, not secrets. Replace local paths and
-load credentials through environment variables or external secret stores.
-
----
-
-## Related Skills
-
-Every MCP gets a generated page:
-
-```text
-mcps/colony/skills.md
-mcps/higgsfield/skills.md
-mcps/medusadocs/skills.md
-mcps/omx/state/skills.md
-```
+## Adding a new MCP
 
-The rules live in `configs/mcp-skill-rules.json`:
+1. Add the server config to `configs/claude.sanitized.json`:
+   ```json
+   "my-server": {
+     "command": "npx",
+     "args": ["my-mcp-server"]
+   }
+   ```
 
-- `source_url` records the upstream repo or hosted endpoint.
-- `homepage`, `package`, `install`, and `health_command` record restore notes.
-- `owners` and `agent_surfaces` record who owns the MCP and where it appears.
-- `include_skills` maps exact skill names.
-- `include_skill_prefixes` maps skill families like `higgsfield-*`.
-- `include_skill_categories` maps whole `recodeee/skills` categories.
-- keyword matching is off by default to avoid noisy false matches.
+2. Create docs at `mcps/my-server/skills.md` describing what tools it exposes.
 
----
+3. Reference it in a profile: `mcps: [my-server]`
 
-## Secrets
+## Security
 
-Do not store secrets in this repo.
+Configs are **sanitized** — no API keys, tokens, or secrets. Environment variables use `<redacted>` placeholders. Real values live in `~/.claude.json` (never committed).
 
-Tracked files may show environment variable names, never values:
+## Related
 
-```text
-COLONY_HOME
-LETSFG_PYTHON
-GITEA_ACCESS_TOKEN
-```
-
-Raw configs are intentionally excluded. Generated JSON and markdown normalize
-local home paths to `~` and redact env values plus secret-loading shell args.
-
-Before pushing, run:
-
-```sh
-./scripts/check-no-secrets.sh
-```
-
-It fails nonzero on local home paths, forbidden secret-like files, and common
-token prefixes.
-
----
-
-```
-   ─── config map, not credential store. keep it boring. ───
-```
+- [cue](https://github.com/opencue/cue) — the profile manager that consumes this registry
+- [resources/skills](../skills/) — the skill library

package/resources/mcps/configs/claude.sanitized.json

@@ -13,6 +13,61 @@
     },
     "cue-tty-watch": {
       "command": "/home/deadpool/Documents/cue/resources/mcps/cue-tty-watch/bin/cue-tty-watch"
+    },
+    "agentshield": {
+      "command": "/home/deadpool/Documents/cue/resources/mcps/agentshield/bin/agentshield"
+    },
+    "x-api": {
+      "command": "npm",
+      "args": [
+        "start"
+      ],
+      "cwd": "${HOME}/.x-api-mcp",
+      "env": {
+        "X_API_KEY": "${X_API_KEY}",
+        "X_API_SECRET": "${X_API_SECRET}",
+        "X_BEARER_TOKEN": "${X_BEARER_TOKEN}",
+        "X_ACCESS_TOKEN": "${X_ACCESS_TOKEN}",
+        "X_ACCESS_TOKEN_SECRET": "${X_ACCESS_TOKEN_SECRET}"
+      }
+    },
+    "trendradar": {
+      "command": "uvx",
+      "args": [
+        "--from",
+        "git+https://github.com/sansan0/TrendRadar.git",
+        "trendradar-mcp"
+      ],
+      "env": {}
+    },
+    "mapi-devdocs": {
+      "type": "http",
+      "url": "https://merchantapi.googleapis.com/devdocs/mcp/"
+    },
+    "mitre-attack": {
+      "command": "/home/deadpool/.local/bin/mitre-attack-mcp",
+      "args": [],
+      "env": {}
+    },
+    "cve-search": {
+      "command": "uv",
+      "args": [
+        "--directory",
+        "/home/deadpool/.config/cue/cache/cve-search_mcp",
+        "run",
+        "main.py"
+      ],
+      "env": {}
+    },
+    "shodan": {
+      "command": "npx",
+      "args": [
+        "-y",
+        "@burtthecoder/mcp-shodan"
+      ],
+      "env": {
+        "SHODAN_API_KEY": "${SHODAN_API_KEY}"
+      }
     }
   },
   "source": "claude",