Skill

Fail

Audit score 45

sf-ai-agentforce-testing

jaganpro/sf-skills

Test Salesforce Agentforce agents with multi-turn API and CLI workflows, coverage analysis, and 100-point scoring.

What is sf-ai-agentforce-testing?

Formal testing skill for Agentforce agents after deployment. Use this when validating multi-turn conversations, topic routing, action invocation, guardrails, and escalation behavior. Supports both Agent Runtime API testing (Track A) and CLI Testing Center workflows (Track B).

Execute multi-turn conversation tests via Agent Runtime API with ECA authentication
Run CLI Testing Center specs with `sf agent test` commands for org-native validation
Generate and analyze test specs covering topics, actions, guardrails, and context preservation
Validate topic routing, action invocation, escalation, and off-topic handling across turns
Classify test failures into root-cause buckets (topic mismatch, action failure, context loss, etc.)
Integrate with fix loops: delegate agent authoring to sf-ai-agentscript, data setup to sf-data, and observability to sf-ai-agentforce-observability

How to install sf-ai-agentforce-testing

npx skills add https://github.com/jaganpro/sf-skills --skill sf-ai-agentforce-testing

Prerequisites

Agent must be published and activated in target org
API v66.0+ (Spring '26) and Agentforce enabled in the org
For Track A (multi-turn API): ECA credentials and Agent Runtime API access configured
For Track B (CLI Testing Center): org must support Agent Testing Center; standard `sf org login web` auth
Target org alias configured in Salesforce CLI

Claude Code

Cursor

Windsurf

Cline

How to use sf-ai-agentforce-testing

1.Gather required context: agent API name, target org alias, testing goal, and publish/activation state
2.Discover the agent in the target org and verify dependencies (Flows, Apex, data)
3.Choose testing track: Track A for multi-turn conversation validation, Track B for CLI Testing Center specs
4.Plan test coverage: main topics, expected actions, guardrails, escalation, and phrasing variations
5.Execute the chosen track using provided scripts under ~/.claude/skills/sf-ai-agentforce-testing/hooks/scripts/
6.Classify failures into root-cause buckets and report results with coverage gaps and next steps
7.Delegate agent fixes to sf-ai-agentscript, republish/reactivate, and rerun focused tests

Use cases

Good for

Smoke-test a newly published Agentforce agent to verify core topics and actions work
Run regression tests after agent updates to ensure topic routing and guardrails still function
Expand test coverage by validating phrasing variations, edge cases, and escalation scenarios
Reproduce reported bugs by replaying multi-turn conversations and analyzing per-turn failures
Validate action invocation chains and context preservation across multiple conversation turns

Who it's for

Salesforce developers testing Agentforce agents post-publish/activate
QA engineers validating agent behavior and coverage before production
Platform engineers integrating agent testing into CI/CD pipelines
Architects analyzing agent test results and planning fix loops

sf-ai-agentforce-testing FAQ

When should I use this skill vs. sf-ai-agentscript?

Use sf-ai-agentforce-testing for validating agents after publish/activate. Use sf-ai-agentscript when you need to edit the agent itself (Agent Script logic, topics, actions). Testing comes after authoring.

What is the difference between Track A and Track B?

Track A (multi-turn API) tests conversation continuity, context preservation, and topic re-matching across turns; requires ECA setup. Track B (CLI Testing Center) runs org-native YAML specs with `sf agent test`; simpler but single-utterance focused.

Do I need ECA credentials for all testing?

No. ECA is only required for Track A (Agent Runtime API multi-turn testing). Track B (CLI Testing Center) and preview workflows use standard `sf org login web` authentication.

What should I do if tests fail?

Classify the failure (topic mismatch, action failure, context loss, etc.), then delegate agent fixes to sf-ai-agentscript if needed. Republish/reactivate and rerun focused tests before full regression.

Can I test an agent before it is published?

No. Testing must happen after the agent is published and activated. Use preview workflows for manual validation during authoring, then switch to formal testing post-publish.

Full instructions (SKILL.md)

Source of truth, from jaganpro/sf-skills.

name: sf-ai-agentforce-testing description: > Agentforce agent testing with dual-track workflow and 100-point scoring. TRIGGER when: user tests Agentforce agents, runs sf agent test commands, creates test specs, validates topic routing, or analyzes agent test coverage. DO NOT TRIGGER when: Apex unit tests (use sf-testing), building agents (use sf-ai-agentforce), or Agent Script DSL (use sf-ai-agentscript). license: MIT compatibility: "Requires API v66.0+ (Spring '26) and Agentforce enabled org" metadata: version: "2.1.0" author: "Jag Valaiyapathy" scoring: "100 points across 7 categories"

sf-ai-agentforce-testing: Agentforce Test Execution & Coverage Analysis

Use this skill when the user needs formal Agentforce testing: multi-turn conversation validation, CLI Testing Center specs, topic/action coverage analysis, preview checks, or a structured test-fix loop after publish.

When This Skill Owns the Task

Use sf-ai-agentforce-testing when the work involves:

sf agent test workflows
multi-turn Agent Runtime API testing
topic routing, action invocation, context preservation, guardrail, or escalation validation
test-spec generation and coverage analysis
post-publish / post-activate test-fix loops

Delegate elsewhere when the user is:

building or editing the agent itself → sf-ai-agentforce or sf-ai-agentscript
running Apex unit tests → sf-testing
creating seed data for actions → sf-data
analyzing session telemetry / STDM traces → sf-ai-agentforce-observability

Core Operating Rules

Testing comes after deploy / publish / activate.
Use multi-turn API testing as the primary path when conversation continuity matters.
Use CLI Testing Center as the secondary path for single-utterance and org-supported test-center workflows.
Interactive and programmatic CLI preview use standard sf org login web authentication; ECA is only required for Agent Runtime API testing, not for live preview.
Fixes to the agent should be delegated to sf-ai-agentscript when Agent Script changes are needed.
Do not use raw curl for OAuth token validation in the ECA flow; use the provided credential tooling.

Script path rule

Use the existing scripts under:

~/.claude/skills/sf-ai-agentforce-testing/hooks/scripts/

These scripts are pre-approved. Do not recreate them.

Required Context to Gather First

Ask for or infer:

agent API name / developer name
target org alias
testing goal: smoke test, regression, coverage expansion, or bug reproduction
whether the agent is already published and activated
whether the org has Agent Testing Center available
whether ECA credentials are available for Agent Runtime API testing

Preflight checks:

discover the agent
confirm publish / activation state
verify dependencies (Flows, Apex, data)
choose testing track

Dual-Track Workflow

Track A — Multi-turn API testing (primary)

Use when you need:

multi-turn conversation testing
topic re-matching validation
context preservation checks
escalation or action-chain analysis across turns

Requires:

ECA / auth setup
agent runtime access

Track B — CLI Testing Center (secondary)

Use when you need:

org-native sf agent test workflows
test spec YAML execution
quick single-utterance validation
CLI-centered CI/CD usage where Testing Center is available

Quick manual path

For manual validation without full formal testing, use preview workflows first, then escalate to Track A or B as needed.

Recommended Workflow

1. Discover and verify

locate the agent in the target org
confirm it is published and activated
confirm required actions / Flows / Apex exist
decide whether Track A or Track B fits the request

2. Plan tests

Cover at least:

main topics
expected actions
guardrails / off-topic handling
escalation behavior
phrasing variation

3. Execute the right track

Track A

validate ECA credentials with the provided tooling
retrieve metadata needed for scenario generation
run multi-turn scenarios with the provided Python scripts
analyze per-turn failures and coverage

Track B

generate or refine a flat YAML test spec
run sf agent test commands
inspect structured results and verbose action output

4. Classify failures

Typical failure buckets:

topic not matched
wrong topic matched
action not invoked
wrong action selected
action invocation failed
context preservation failure
guardrail failure
escalation failure

5. Run fix loop

When failures imply agent-authoring issues:

delegate fixes to sf-ai-agentscript
re-publish / re-activate if needed
re-run focused tests before full regression

Testing Guardrails

Never skip these:

test only after publish/activate
include harmful / off-topic / refusal scenarios
use multiple phrasings per important topic
clean up sessions after API tests
keep swarm execution small and controlled

Avoid these anti-patterns:

testing unpublished agents
treating one happy-path utterance as coverage
storing ECA secrets in repo files
debugging auth with brittle shell-expanded curl commands
changing both tests and agent simultaneously without isolating the cause

Output Format

When finishing a run, report in this order:

Test track used
What was executed
Pass/fail summary
Coverage gaps
Root-cause themes
Recommended fix loop / next test step

Suggested shape:

Agent: <name>
Track: Multi-turn API | CLI Testing Center | Preview
Executed: <specs / scenarios / turns>
Result: <passed / partial / failed>
Coverage: <topics, actions, guardrails, context>
Issues: <highest-signal failures>
Next step: <fix, republish, rerun, or expand coverage>

Cross-Skill Integration

Need	Delegate to	Reason
fix Agent Script logic	sf-ai-agentscript	authoring and deterministic fix loops
create test data	sf-data	action-ready data setup
fix Flow-backed actions	sf-flow	Flow repair
fix Apex-backed actions	sf-apex	Apex repair
set up ECA / OAuth for Agent Runtime API	sf-connected-apps	auth and app configuration
analyze session telemetry	sf-ai-agentforce-observability	STDM / trace analysis

Reference Map

Start here

Execution / auth

Coverage / fix loops

Advanced / specialized

Templates / assets

Score Guide

Score	Meaning
90+	production-ready test confidence
80–89	strong coverage with minor gaps
70–79	acceptable but coverage expansion recommended
60–69	partial validation only
< 60	insufficient confidence; block release