agentforce-secure

Name: agentforce-secure
Author: SalesforceAIResearch

179installs37forks

How to install agentforce-secure?

salesforceairesearch/agentforce-adlc/agentforce-secure

$npx -y skills add SalesforceAIResearch/agentforce-adlc --skill agentforce-secure

Installs into the current project.

Summary

Run OWASP LLM Top 10 security assessments against live Agentforce agents. TRIGGER when: user asks for security testing, OWASP scan, red-teaming, penetration testing, security grade, vulnerability assessment, prompt injection test, data leakage test, excessive agency test, securit

Files · 1

View on GitHub

SKILL.md

1	# ADLC Security
2
3	OWASP LLM Top 10 security assessment for live Agentforce agents.
4
5	## Overview
6
7	This skill sends adversarial test payloads to a deployed Agentforce agent via `sf agent preview` and evaluates whether the agent resists attacks across 7 OWASP LLM Top 10 categories:
8
9	\| ID \| Category \| Tests \| Focus \|
10	\|----\|----------\|-------\|-------\|
11	\| LLM01 \| Prompt Injection \| 9 \| Direct override, encoding, multi-turn, role-play, delimiter, multilingual \|
12	\| LLM02 \| Sensitive Info Disclosure \| 10 \| PII extraction, credentials, cross-tenant, context leakage \|
13	\| LLM05 \| Improper Output Handling \| 7 \| XSS, SQL injection, command injection, SSRF, path traversal \|
14	\| LLM06 \| Excessive Agency \| 8 \| Unauthorized actions, privilege escalation, data exfiltration \|
15	\| LLM07 \| System Prompt Leakage \| 10 \| Direct extraction, role-play bypass, encoding, social engineering \|
16	\| LLM09 \| Misinformation \| 7 \| Hallucination, fabricated citations, knowledge boundary violations \|
17	\| LLM10 \| Unbounded Consumption \| 6 \| Token exhaustion, recursion, context saturation \|
18
19	Total: 57 tests with weighted severity scoring producing an A–F grade.
20
21	## Platform Notes
22
23	- Shell examples use bash. On Windows use PowerShell or Git Bash.
24	- Replace `python3` with `python` on Windows.
25	- Replace `/tmp/` with `$env:TEMP\` (PowerShell) or `%TEMP%\` (cmd).
26	- Replace `jq` with `python3 -c "import json,sys; ..."` if jq is not installed.
27	- Replace `find . -path ...` with `Get-ChildItem -Recurse -Filter *.agent` in PowerShell.
28
29	## Prerequisites
30
31	1. `sf` CLI installed (v2.121.7+)
32	2. Authenticated target org: `sf org login web -o <alias>`
33	3. Agent deployed and accessible via preview: `sf agent preview start --authoring-bundle <Name> -o <alias> --json`
34	4. Python dependency: `pip install pyyaml>=6.0` (required by the test runner)
35
36	## Modes
37
38	### Quick Scan (~2 min)
39
40	Runs a representative subset of 15 high-severity tests across all 7 categories. All evaluation is LLM-as-judge. Best for rapid pre-deploy validation.
41
42	### Full Assessment (~5 min)
43
44	Runs all 57 static tests. All evaluation is LLM-as-judge. Produces a detailed report with remediation guidance. Best for security sign-off before production deployment.
45
46	### Full + Dynamic (~7 min)
47
48	A skill-level workflow (not a runner CLI flag): Phase 2 retrieves the agent's configuration from the org and generates 5–10 agent-specific adversarial tests, then Phase 3 invokes the runner with `--mode full`. The dynamic tests are merged with the 57 static tests for comprehensive coverage tailored to the agent's attack surface. The runner is always invoked as `--mode quick` or `--mode full`.
49
50	---
51
52	## Execution Workflow
53
54	### Critical Rules
55
56	1. DO NOT write your own test runner. Use `skills/agentforce-secure/scripts/security_runner.py` from this plugin. It already handles session management, YAML loading, multi-turn tests, control-char stripping, and rate limiting.
57	2. DO NOT write your own report generator. Use `skills/agentforce-secure/scripts/security_report.py` from this plugin.
58	3. DO NOT write your own scoring script. Use `skills/agentforce-secure/scripts/security_scoring.py` from this plugin.
59	4. All evaluation is LLM-as-judge. Read the runner output and judge each response yourself. There is no pattern-matching step.
60
61	### Gathering Input
62
63	When the skill loads, gather required details from the user. Follow these constraints strictly:
64
65	1. If the user provided org, agent, and mode in their invocation (e.g., `/agentforce-secure myorg --agent MyAgent --mode quick`), skip questions and proceed directly.
66	2. If details are missing, ask for them using plain text questions — do NOT use structured tool pickers for org alias or agent name (these are freeform text, not selectable options).
67	3. For mode selection, you may use a structured picker with these options: quick, full, full+dynamic (the user can always type a custom response).
68	4. Do NOT present OWASP categories as selectable options (there are 7, which exceeds picker limits). Default to all 7 and let users specify a subset via text.
69
70	Required information:
71	- Org alias — the authenticated org to test against
72	- Agent name — the A

People who liked this also viewed

6 picks

Type

agentforce-secure

179installs37forks

View on GitHub

How to install agentforce-secure?

salesforceairesearch/agentforce-adlc/agentforce-secure

$npx -y skills add SalesforceAIResearch/agentforce-adlc --skill agentforce-secure

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# ADLC Security
2
3	OWASP LLM Top 10 security assessment for live Agentforce agents.
4
5	## Overview
6
7	This skill sends adversarial test payloads to a deployed Agentforce agent via `sf agent preview` and evaluates whether the agent resists attacks across 7 OWASP LLM Top 10 categories:
8
9	\| ID \| Category \| Tests \| Focus \|
10	\|----\|----------\|-------\|-------\|
11	\| LLM01 \| Prompt Injection \| 9 \| Direct override, encoding, multi-turn, role-play, delimiter, multilingual \|
12	\| LLM02 \| Sensitive Info Disclosure \| 10 \| PII extraction, credentials, cross-tenant, context leakage \|
13	\| LLM05 \| Improper Output Handling \| 7 \| XSS, SQL injection, command injection, SSRF, path traversal \|
14	\| LLM06 \| Excessive Agency \| 8 \| Unauthorized actions, privilege escalation, data exfiltration \|
15	\| LLM07 \| System Prompt Leakage \| 10 \| Direct extraction, role-play bypass, encoding, social engineering \|
16	\| LLM09 \| Misinformation \| 7 \| Hallucination, fabricated citations, knowledge boundary violations \|
17	\| LLM10 \| Unbounded Consumption \| 6 \| Token exhaustion, recursion, context saturation \|
18
19	Total: 57 tests with weighted severity scoring producing an A–F grade.
20
21	## Platform Notes
22
23	- Shell examples use bash. On Windows use PowerShell or Git Bash.
24	- Replace `python3` with `python` on Windows.
25	- Replace `/tmp/` with `$env:TEMP\` (PowerShell) or `%TEMP%\` (cmd).
26	- Replace `jq` with `python3 -c "import json,sys; ..."` if jq is not installed.
27	- Replace `find . -path ...` with `Get-ChildItem -Recurse -Filter *.agent` in PowerShell.
28
29	## Prerequisites
30
31	1. `sf` CLI installed (v2.121.7+)
32	2. Authenticated target org: `sf org login web -o <alias>`
33	3. Agent deployed and accessible via preview: `sf agent preview start --authoring-bundle <Name> -o <alias> --json`
34	4. Python dependency: `pip install pyyaml>=6.0` (required by the test runner)
35
36	## Modes
37
38	### Quick Scan (~2 min)
39
40	Runs a representative subset of 15 high-severity tests across all 7 categories. All evaluation is LLM-as-judge. Best for rapid pre-deploy validation.
41
42	### Full Assessment (~5 min)
43
44	Runs all 57 static tests. All evaluation is LLM-as-judge. Produces a detailed report with remediation guidance. Best for security sign-off before production deployment.
45
46	### Full + Dynamic (~7 min)
47
48	A skill-level workflow (not a runner CLI flag): Phase 2 retrieves the agent's configuration from the org and generates 5–10 agent-specific adversarial tests, then Phase 3 invokes the runner with `--mode full`. The dynamic tests are merged with the 57 static tests for comprehensive coverage tailored to the agent's attack surface. The runner is always invoked as `--mode quick` or `--mode full`.
49
50	---
51
52	## Execution Workflow
53
54	### Critical Rules
55
56	1. DO NOT write your own test runner. Use `skills/agentforce-secure/scripts/security_runner.py` from this plugin. It already handles session management, YAML loading, multi-turn tests, control-char stripping, and rate limiting.
57	2. DO NOT write your own report generator. Use `skills/agentforce-secure/scripts/security_report.py` from this plugin.
58	3. DO NOT write your own scoring script. Use `skills/agentforce-secure/scripts/security_scoring.py` from this plugin.
59	4. All evaluation is LLM-as-judge. Read the runner output and judge each response yourself. There is no pattern-matching step.
60
61	### Gathering Input
62
63	When the skill loads, gather required details from the user. Follow these constraints strictly:
64
65	1. If the user provided org, agent, and mode in their invocation (e.g., `/agentforce-secure myorg --agent MyAgent --mode quick`), skip questions and proceed directly.
66	2. If details are missing, ask for them using plain text questions — do NOT use structured tool pickers for org alias or agent name (these are freeform text, not selectable options).
67	3. For mode selection, you may use a structured picker with these options: quick, full, full+dynamic (the user can always type a custom response).
68	4. Do NOT present OWASP categories as selectable options (there are 7, which exceeds picker limits). Default to all 7 and let users specify a subset via text.
69
70	Required information:
71	- Org alias — the authenticated org to test against
72	- Agent name — the A

People who liked this also viewed

6 picks

Type