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/llm-controlled-actions.agent
View on GitHub# LLM-Controlled Actions Pattern
# =================================
#
# A complete, read-only catalog agent that contrasts:
# - deterministic action execution in a lifecycle hook; and
# - model-chosen action execution from reasoning.actions.
#
# Use deterministic `run` only when execution is required whenever its
# containing hook is reached. Put optional, intent-dependent actions in
# `reasoning.actions` so the model can choose whether to invoke them.
#
# This example intentionally exposes no action that changes money, records,
# access, commitments, or other external state. UI confirmation and prose
# instructions alone are not sufficient guards for consequential actions.
system:
instructions: |
You are a public product-catalog assistant. Help users discover products
and check current availability. Use the read-only actions before making
claims about catalog matches or availability. Treat action results as
the source of truth, and say when a lookup returns no result or fails.
Ask only for information required by the requested lookup.
messages:
welcome: "Hello! I can help you search the catalog or check product availability."
error: "I couldn't retrieve the requested catalog information. Please try again."
config:
developer_name: "LLM_Controlled_Actions"
agent_label: "Catalog Action Patterns"
description: "Demonstrates deterministic hooks and model-chosen read-only actions"
access:
default_agent_user: "agent@company.salesforce.com"
variables:
service_notice: mutable string = ""
description: "Trusted notice refreshed before every turn; the hook owns and overwrites it, and reasoning reads only the current-turn value"
language:
default_locale: "en_US"
start_agent catalog_assistance:
label: "Catalog Assistance"
description: "Searches a public catalog and checks current product availability"
actions:
get_service_notice:
description: "Returns the current public catalog service notice for this locale"
inputs:
locale: string
description: "Locale for the service notice"
outputs:
notice: string
description: "Current service notice, or an empty string when none applies"
target: "flow://Get_Catalog_Service_Notice"
search_catalog:
description: "Searches the public catalog; choose this for product discovery, not for current availability"
inputs:
query: string
description: "Product name, category, or features requested by the user"
outputs:
matches: string
description: "Current catalog matches returned by the search"
is_displayable: True
target: "flow://Search_Public_Catalog"
check_availability:
description: "Checks current public availability; choose this when the user asks whether or where a product is available"
inputs:
product: string
description: "Product name or identifier to check"
postal_code: string
description: "Postal code where availability should be checked"
outputs:
availability: string
description: "Current availability returned by the inventory service"
is_displayable: True
target: "flow://Check_Public_Availability"
# service_notice lifecycle:
# - owner: catalog_assistance
# - writer: this before_reasoning hook and get_service_notice output
# - reader: the current turn's reasoning instructions
# - reset: empty before each refresh
# - expiry: the next reasoning turn
# - correction: the next trusted refresh replaces the current notice
# - cancellation: no reset is needed because the notice neither gates nor
# changes an action
# Deterministic refresh is justified by external freshness ordering.
before_reasoning:
set @variables.service_notice = ""
run @actions.get_service_notice
with locale="en_US"
set @variables.service_notice = @outputs.notice
reasoning:
instructions: ->
| Help with the user's current public-catalog request.
The service notice refreshed for this turn is:
{!@variables.service_notice}
Mention that notice only when it is relevant.
Use catalog_search for product discovery.
Use availability_lookup only for current availability.
If a required lookup value is absent, ask for that value instead
of invoking an action with a guess.
# Model-chosen: these read-only actions are available to reasoning, but
# execute only when the user's current request calls for them.
actions:
catalog_search: @actions.search_catalog
with query=...
availability_lookup: @actions.check_availability
with product=...
with postal_code=...
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 |