iii-worker-manager
WebSocket listener that SDK workers connect to. Use to expose the engine to trusted (internal) and untrusted (RBAC-protected) workers, gate function and trigger registration, and audit or transform every invocation through a middleware function. Mounts channel WebSocket endpoints on every listener.
What this skill does
# iii-worker-manager
The `iii-worker-manager` is a mandatory engine worker that opens WebSocket listeners. The first entry in `iii-config.yaml` sets the main engine port (`49134` by default) for trusted internal traffic; additional entries open separate listeners with their own RBAC, middleware, and registration hooks. Channels are mounted on every listener at `/ws/channels/{channel_id}` — RBAC ports keep `engine::channels::create` always-allowed via the infrastructure carve-out so SDK `createChannel()` works without changes.
## When to Use
- Exposing the engine to an untrusted network — add a second `iii-worker-manager` entry with an `rbac` block and an `auth_function_id` instead of opening the main engine port.
- Restricting which functions a connected worker can invoke — combine `expose_functions` filters (operator-side) with `forbidden_functions` from `AuthResult` (per-session, hard deny).
- Auditing, rate-limiting, or enriching every invocation — set `middleware_function_id` on the listener; the middleware decides whether to call the target and what to return.
- Per-tenant or per-session namespace isolation — return `function_registration_prefix` from the auth function so every function/trigger this session registers is transparently prefixed without the worker code knowing.
- Gating dynamic registration — wire `on_function_registration_function_id`, `on_trigger_registration_function_id`, or `on_trigger_type_registration_function_id` to validate or rewrite registrations.
## Boundaries
- The infrastructure carve-out (`engine::channels::create`, `engine::workers::register`, `engine::log::*`, `engine::baggage::*`) is always allowed on RBAC listeners regardless of `expose_functions`. Adding one of those IDs to `forbidden_functions` denies it but logs a warning — workers may behave unpredictably (broken setup, lost logs, missing context).
- The first `iii-worker-manager` entry is the main engine port and should remain internal. Only RBAC-protected listeners belong on external networks.
- The middleware is **not** a pre-handler — it must invoke the target function itself (typically via `iii.trigger`) and return its result. Returning early without invoking simply skips the call.
- `forbidden_functions` from the auth result wins over both `allowed_functions` and `expose_functions`. There is no way for a session to override an operator's deny list.
- Registration hooks return mapped fields or **throw** to deny. Omitted result fields keep their original values; returning `{}` is a no-op (allow as-is).
- Triggers and functions registered through an RBAC session are scoped to that session and cleaned up automatically on disconnect.
## Configuration
Two listeners — one internal, one external with RBAC:
```yaml
workers:
- name: iii-worker-manager
config:
port: 49134
- name: iii-worker-manager
config:
host: 0.0.0.0
port: 49135
middleware_function_id: my-project::middleware
rbac:
auth_function_id: my-project::auth
on_function_registration_function_id: my-project::on-fn-reg
on_trigger_registration_function_id: my-project::on-trig-reg
on_trigger_type_registration_function_id: my-project::on-trig-type-reg
expose_functions:
- match("api::*")
- match("*::public")
- metadata:
public: true
```
## Auth function
Receives `AuthInput { headers, query_params, ip_address }`, returns `AuthResult`. Throw to reject the connection.
```typescript
import type { AuthInput, AuthResult } from 'iii-sdk'
iii.registerFunction(
'my-project::auth',
async (input: AuthInput): Promise<AuthResult> => {
const token = input.headers?.['authorization']?.replace(/^Bearer\s+/i, '')
if (!token) throw new Error('Missing credentials')
const user = await validateToken(token)
return {
allowed_functions: [],
forbidden_functions: user.role === 'readonly'
? ['api::users::delete', 'api::users::update']
: [],
allowed_trigger_types: user.role === 'admin' ? ['cron', 'webhook'] : undefined,
allow_trigger_type_registration: user.role === 'admin',
function_registration_prefix: `tenant-${user.tenant_id}`,
context: { user_id: user.id, role: user.role, tenant_id: user.tenant_id },
}
},
)
```
`AuthResult` defaults: `allow_function_registration: true`, `allow_trigger_type_registration: false`, `allowed_trigger_types: undefined` (permissive — every type allowed). Setting `function_registration_prefix` makes the engine transparently rewrite IDs in both directions; the worker SDK never sees the prefix.
## Middleware function
Runs on every invocation through this listener. Receives `MiddlewareFunctionInput { function_id, payload, action, context }`. Must call the target function itself.
```typescript
import type { MiddlewareFunctionInput } from 'iii-sdk'
iii.registerFunction(
'my-project::middleware',
async (input: MiddlewareFunctionInput) => {
console.log(`[audit] user=${input.context.user_id} → ${input.function_id}`)
return iii.trigger({
function_id: input.function_id,
payload: {
...input.payload,
_caller_id: input.context.user_id,
_caller_role: input.context.role,
},
})
},
)
```
Skipping the `iii.trigger` call short-circuits the request — useful for rate-limit rejections (return a structured error envelope instead of invoking the target).
## Registration hooks
Each hook receives the registration details plus `AuthResult.context`. Return mapped fields or throw to deny. Omitted fields keep their original value.
```typescript
import type {
OnFunctionRegistrationInput,
OnTriggerRegistrationInput,
OnTriggerTypeRegistrationInput,
} from 'iii-sdk'
iii.registerFunction(
'my-project::on-fn-reg',
async (input: OnFunctionRegistrationInput) => {
if (input.function_id.startsWith('internal::')) {
throw new Error('Cannot register internal functions')
}
return { function_id: input.function_id }
},
)
iii.registerFunction(
'my-project::on-trig-reg',
async (input: OnTriggerRegistrationInput) => {
const role = input.context.role as string
if (!input.function_id.startsWith(`${role}::`)) {
throw new Error('Function ID must be prefixed with the role')
}
return { function_id: `${role}::${input.function_id}` }
},
)
iii.registerFunction(
'my-project::on-trig-type-reg',
async (input: OnTriggerTypeRegistrationInput) => {
if (input.context.role !== 'admin') {
throw new Error('Only admins can register trigger types')
}
return {}
},
)
```
A worker can register a trigger type only when `allow_trigger_type_registration: true` AND (if configured) the hook returns a result. A worker can register a trigger only when its `trigger_type` is in `allowed_trigger_types` (or the field is omitted) AND the hook returns a result.
## Function filters
`expose_functions` accepts wildcard and metadata filters; multiple filters are OR'd. Metadata filters AND across keys.
```yaml
expose_functions:
- match("api::*") # ID glob: * matches any chars
- match("*::public") # suffix match
- match("api::*::read") # multi-segment match
- metadata: # all keys must match
public: true
tier: free
- metadata:
name: match("*public*") # metadata values can be wildcards too
```
A function is exposed if **any** filter matches. Empty `expose_functions` = no functions exposed (other than the infrastructure carve-out).
## Channels on RBAC ports
Channels are mounted at `/ws/channels/{channel_id}` on the same port the worker connected through. `engine::channels::create` is always allowed (infrastructure carve-out), so RBAC listeners can hand out channel refs even with empty `expose_functions`. SDK `createChannel()` works unmodified — channel data flows through whichever listener the worker is on, never the main engine port. Access to a channeRelated in Backend & APIs
jfrog
IncludedInteract with the JFrog Platform via the JFrog CLI and REST/GraphQL APIs. Use this skill when the user wants to manage Artifactory repositories, upload or download artifacts, manage builds, configure permissions, manage users and groups, work with access tokens, configure JFrog CLI servers, search artifacts, manage properties, set up replication, manage JFrog Projects, run security audits or scans, look up CVE details, query exposures scan results from JFrog Advanced Security, manage release bundles and lifecycle operations, aggregate or export platform data, or perform any JFrog Platform administration task. Also use when the user mentions jf, jfrog, artifactory, xray, distribution, evidence, apptrust, onemodel, graphql, workers, mission control, curation, advanced security, exposures, or any JFrog product name.
cupynumeric-migration-readiness
IncludedPre-migration readiness assessor for porting NumPy to cuPyNumeric. Use BEFORE substantial porting work begins when the user asks whether code will scale on GPU, whether they should migrate to cuPyNumeric, which NumPy patterns transfer cleanly, what must be refactored before porting, or mentions pre-port assessment, scaling analysis, or refactor planning. Inspect the user's source code, look up NumPy usage, cross-reference the cuPyNumeric API support manifest, and distinguish distributed-scaling-friendly patterns from blockers such as unsupported APIs, scalar synchronization, host round-trips, Python/object-heavy control flow, shape/data-dependent branching, and in-place mutation hazards. Produce a verdict of READY, LIGHT REFACTOR, SIGNIFICANT REFACTOR, or NOT RECOMMENDED, with concrete refactor pointers.
alibabacloud-data-agent-skill
IncludedInvoke Alibaba Cloud Apsara Data Agent for Analytics via CLI to perform natural language-driven data analysis on enterprise databases. Data Agent for Analytics is an intelligent data analysis agent developed by Alibaba Cloud Database team for enterprise users. It automatically completes requirement analysis, data understanding, analysis insights, and report generation based on natural language descriptions. This tool supports: discovering data resources (instances/databases/tables) managed in DMS, initiating query or deep analysis sessions, real-time progress tracking, and retrieving analysis conclusions and generated reports. Use this Skill when users need to query databases, analyze data trends, generate data reports, ask questions in natural language, or mention "Data Agent", "data analysis", "database query", "SQL analysis", "data insights".
token-optimizer
IncludedReduce OpenClaw token usage and API costs through smart model routing, heartbeat optimization, budget tracking, and native 2026.2.15 features (session pruning, bootstrap size limits, cache TTL alignment). Use when token costs are high, API rate limits are being hit, or hosting multiple agents at scale. The 4 executable scripts (context_optimizer, model_router, heartbeat_optimizer, token_tracker) are local-only — no network requests, no subprocess calls, no system modifications. Reference files (PROVIDERS.md, config-patches.json) document optional multi-provider strategies that require external API keys and network access if you choose to use them. See SECURITY.md for full breakdown.
resend-cli
IncludedUse this skill when the task is specifically about operating Resend from an AI agent, terminal session, or CI job via the official resend CLI: installing/authenticating the CLI, sending/listing/updating/cancelling emails, batch sends, domains and DNS, webhooks and local listeners, inbound receiving, contacts, topics, segments, broadcasts, templates, API keys, profiles, or debugging Resend CLI/API failures. Trigger on mentions of Resend CLI, `resend`, `resend doctor`, `resend emails send`, `resend domains`, `resend webhooks listen`, `resend emails receiving`, or agent-friendly terminal automation.
alibabacloud-odps-maxframe-coding
IncludedUse this skill for MaxFrame SDK development and documentation navigation on Alibaba Cloud MaxCompute (ODPS). Helps answer MaxFrame API, concept, official example, and supported pandas API questions; create data processing programs; read/write MaxCompute tables; debug jobs (remote or local); and build custom DPE runtime images. Trigger when users mention MaxFrame, MaxCompute with MaxFrame, ODPS table processing, DPE runtime, MaxFrame docs/examples, DataFrame/Tensor operations, or GPU runtime setup. Works for both English and Chinese queries about Alibaba Cloud data processing with MaxFrame.