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.
references/optimization-pattern-3-reference-syntax.md
View on GitHub# Optimization Pattern 3: Fix Variable and Action Reference Syntax
## Detection Logic
Scan all `reasoning.instructions` blocks for two categories of issues:
### 1. Incorrect reference syntax
- Model instructions require a stored value but contain bare `@variables.X`,
which is only literal text
- Actions/tools mentioned by name without using `{!@actions.X}` syntax
### 2. Missing action references
- Scan all actions in `reasoning.actions` and their definitions in `actions:` section
- Identify what use case each action serves based on its name, description, inputs/outputs
- Check if `reasoning.instructions` mentions those use cases without explicit `{!@actions.X}` references
- Match phrases like "retrieve details", "get contact", "update status", "look up", "search for" to action names
## How to Fix
Classify each reference before editing:
1. If the model must read the current value, change bare `@variables.X` to
`{!@variables.X}`.
2. If the text merely names an action parameter or state concept, keep it
literal. For example, `Set next_destination to "self_service"` does not need
the current value of `next_destination`.
3. If the text performs an exact comparison over a trusted value, do not stop
at interpolation. Move the comparison into an AgentScript `if` when the
branch must be deterministic.
4. Add `{!@actions.X}` when model instructions need to identify an available
action by its use case.
Do not auto-fix every interpolation lint hint. The question is whether the
model needs the value, not whether a variable-like name appears in prose.
## Example
**Before:**
```agentscript
subagent CustomerService:
reasoning:
instructions: ->
| Address user as @variables.userName. Help customers by looking up their account information and answering their questions.
actions:
GetAccount: @actions.GetCustomerAccount
with email = @variables.customerEmail
AnswerQuestion: @actions.AnswerWithKnowledge
with query = ...
actions:
GetCustomerAccount:
description: "Retrieves customer account details"
inputs:
"email": string
outputs:
"accountInfo": object
AnswerWithKnowledge:
description: "Answers questions using knowledge base"
inputs:
"query": string
```
**After:**
```agentscript
subagent CustomerService:
reasoning:
instructions: ->
| Address user as {!@variables.userName}. Help customers by looking up their account information with {!@actions.GetAccount} and answering their questions with {!@actions.AnswerQuestion}.
actions:
GetAccount: @actions.GetCustomerAccount
with email = @variables.customerEmail
AnswerQuestion: @actions.AnswerWithKnowledge
with query = ...
actions:
GetCustomerAccount:
description: "Retrieves customer account details"
inputs:
"email": string
outputs:
"accountInfo": object
AnswerWithKnowledge:
description: "Answers questions using knowledge base"
inputs:
"query": string
```
**Key improvements:**
- Fixed variable reference: `@variables.userName` → `{!@variables.userName}`
- Matched action use cases to instructions:
- "looking up their account information" → `GetCustomerAccount` → added `{!@actions.GetAccount}`
- "answering their questions" → `AnswerWithKnowledge` → added `{!@actions.AnswerQuestion}`
- Explicit action references improve the LLM's ability to select the correct action at runtime
## Interpolation Is Not Control Flow
```agentscript
# Still wrong when this must be an exact runtime branch
| If {!@variables.customer_tier} == "vip", use priority support.
# Runtime-owned exact fact
if @variables.customer_tier == "vip":
| Offer the priority-support options available to this customer.
# Model-owned semantic judgment
| If the user is asking for a human rather than product guidance, use
{!@actions.escalate_to_support}.
```
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 |