forcedotcom/sf-skills / plugins/builder/agentforce-adlc/skills/agentforce-generate
agentforce-generate
FlaggedBuild, modify, audit, repair, optimize, debug, and deploy agents with Agentforce Agent Script. TRIGGER when: user creates, reviews, or changes .agent files or aiAuthoringBundle metadata; asks to fix AgentScript, audit an existing agent, run an AgentScript health check, common-pitfall review, or baseline-versus-candidate repair loop; changes a response, action, subagent, route, state flow, or Agent Spec; migrates or upgrades a legacy GenAiPlannerBundle Agentforce agent (NOT an Einstein Bot) to an aiAuthoringBundle (Agent Script) agent; previews, debugs, deploys, publishes, or tests agents; uses sf agent generate/preview/publish/test; or manages Agentforce MCP servers, tools, assets, or authentication. DO NOT TRIGGER when: Apex, Flow, Prompt Template, Experience Cloud, or general Salesforce CLI work is unrelated to Agent Script; the source is an Einstein Bot (BotDefinition/BotVersion) — use agentforce-bot-upgrade; or the primary input is a production session or trace ID rather than an agent artifact.
assets/patterns/advanced-input-bindings.agent
View on GitHub# Advanced Input Bindings Pattern
# Demonstrates all parameter binding techniques for Agent Script actions
#
# ★ When To Use This Pattern:
# - Learning different ways to pass values to actions
# - Combining LLM slot filling with variable binding
# - Chaining outputs between multiple actions
# - Complex multi-input action scenarios
#
# ★ Key Insight:
# - `...` (ellipsis) = LLM extracts value from conversation (slot filling)
# - `"value"` = Fixed constant value
# - `@variables.x` = Value from stored state
# - `@outputs.x` = Value from previous action's output
#
# ★ Common Use Cases:
# - Order lookup with user-provided order ID
# - Multi-step workflows with data passing
# - Conditional parameter binding
#
# This is a COMPLETE template - customize action targets and metadata.
system:
instructions: |
You are an order-processing assistant. Use slot filling for safe values
present in the conversation, literals only for real constants, and
stored state only for canonical action results required later.
messages:
welcome: "Hello! I can demonstrate safe order input bindings."
error: "I couldn't process that request."
access:
default_agent_user: "agent@company.salesforce.com"
config:
developer_name: "Advanced_Input_Bindings_Agent"
agent_label: "Advanced Input Bindings"
description: "Demonstrates slot, literal, canonical-state, and action-output bindings"
variables:
current_account_id: mutable string = ""
description: "Canonical account ID returned by account resolution and required by later actions"
language:
default_locale: "en_US"
start_agent order_processing:
label: "Order Processing"
description: "Demonstrates advanced input binding patterns"
actions:
# Action with multiple input types
process_order:
description: "Process an order with various input methods"
inputs:
order_id: string
description: "The order ID to process"
amount: object
description: "Transaction amount in USD"
complex_data_type_name: "lightning__doubleType"
account_id: string
description: "Customer account ID"
outputs:
confirmation_number: string
description: "Order confirmation number"
processed_amount: object
description: "Final processed amount"
complex_data_type_name: "lightning__doubleType"
target: "flow://Process_Order"
get_account:
description: "Look up account details"
inputs:
account_id: string
description: "Account ID to look up"
outputs:
account_name: string
description: "Account name"
credit_limit: object
description: "Available credit"
complex_data_type_name: "lightning__doubleType"
target: "flow://Get_Account_Details"
resolve_account:
description: "Resolve the account named by the user to a canonical account ID"
inputs:
account_reference: string
description: "Account name or reference from the conversation"
outputs:
account_id: string
description: "Canonical account ID"
target: "flow://Resolve_Account"
send_notification:
description: "Send order notification"
inputs:
confirmation_number: string
description: "Order confirmation to include"
recipient_account: string
description: "Account to notify"
outputs:
sent: boolean
description: "Whether notification was sent"
target: "flow://Send_Order_Notification"
reasoning:
instructions: ->
| Help the customer process their order.
INPUT BINDING EXAMPLES:
1. Slot filling (...) - LLM extracts from conversation
2. Fixed values - Always use a constant
3. Stored state - Use the available value
4. Output chaining - Use results from previous action
Ask for order details and process appropriately.
actions:
# Persist only a canonical action output that later bindings consume.
resolve_account: @actions.resolve_account
with account_reference=...
set @variables.current_account_id = @outputs.account_id
# ★ PATTERN 1: Slot Filling (LLM extracts from conversation)
# User says "Process order ORD-12345" -> LLM fills order_id="ORD-12345"
slot_fill_lookup: @actions.process_order
with order_id=...
with amount=...
with account_id=...
# ★ PATTERN 2: Fixed Value (constant)
# Always uses the same account for default lookups
fixed_account: @actions.get_account
with account_id="001DEFAULT000001"
# ★ PATTERN 3: Variable Binding (from stored state)
# Uses the canonical account ID returned by resolve_account
variable_binding: @actions.get_account
with account_id=@variables.current_account_id
available when @variables.current_account_id != ""
# ★ PATTERN 4: Output Chaining (from previous action)
# Uses confirmation_number from process_order to send notification
chained_output: @actions.process_order
with order_id=...
with amount=...
with account_id=@variables.current_account_id
run @actions.send_notification
with confirmation_number=@outputs.confirmation_number
with recipient_account=@variables.current_account_id
available when @variables.current_account_id != ""
# ★ PATTERN 5: Mixed Binding (combining patterns)
# Some inputs from LLM, some from variables, some fixed
mixed_binding: @actions.process_order
with order_id=... # LLM slot fills
with amount=... # LLM slot fills
with account_id="001INTERNAL00001" # Fixed value
# ★ Insight: Binding Pattern Decision Tree
#
# Need the LLM to extract from conversation? -> Use `...`
# Value is always the same constant? -> Use "value"
# Value was captured in a previous step? -> Use @variables.x
# Value comes from another action's output? -> Use @outputs.x (in callback)
#
# ★ Common Mistake: Using @outputs.x outside of a `run` callback
# @outputs.x is only available inside the action's callback chain
# Store important outputs in @variables for use elsewhere
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |