audit-deployed-package

audit-deployed-package is a skill for Claude Code from runverdict/sf-security-review-toolkit. It costs 153 tokens per session (8,486 once invoked), scanned A, original, Apache-2.0.

A procedure for checking a Salesforce package after it has been installed into a temporary organization. It tests the installed artifact's behavior, rather than only reading the package's source files.

In plain words
What is it for?
Verifying subscriber-facing permissions, install-time grants, CRUD and field-level security behavior, configured callouts, and uninstall cleanup. Its findings are added to the same security-audit record as the source review.
Why use it?
Source inspection cannot reveal every install-time permission change, runtime data-access issue, callout configuration problem, or uninstall cleanup issue. This audit adds evidence from a real installation in a throwaway Salesforce organization.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: mentions subagents; names the AskUserQuestion tool.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the sf-security-review-toolkit plugin — 14 skills, 1 hook shipped together

Good fit Verifying subscriber-facing permissions, install-time grants, CRUD and field-level security behavior, configured callouts, and uninstall cleanup. Its findings are added to the same security-audit record as the source review.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add runverdict/sf-security-review-toolkit
Claude Code
/plugin install sf-security-review-toolkit

Made for: Claude Code.

Or install sf-security-review-toolkit, the plugin that ships this one along with the rest of its 14 skills, 1 hook.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for audit-deployed-package

README.md
[![agentmods](https://agentmods.dev/badge/skills/runverdict/sf-security-review-toolkit/audit-deployed-package/github.svg)](https://agentmods.dev/skills/runverdict/sf-security-review-toolkit/audit-deployed-package)
Your own site
<a href="https://agentmods.dev/skills/runverdict/sf-security-review-toolkit/audit-deployed-package"><img src="https://agentmods.dev/badge/skills/runverdict/sf-security-review-toolkit/audit-deployed-package/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for audit-deployed-package

Your own site · 80×15
<a href="https://agentmods.dev/skills/runverdict/sf-security-review-toolkit/audit-deployed-package"><img src="https://agentmods.dev/badge/skills/runverdict/sf-security-review-toolkit/audit-deployed-package.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 153 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 8,486 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00153 $0.08486
Opus 5 $0.00077 $0.04243
Sonnet 5 $0.00031 $0.01697
Haiku 4.5 $0.00015 $0.00849

Measured 9d ago against content hash 1953d4266a4f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

audit-deployed-package scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/audit-deployed-package/SKILL.md · 512 lines

How it starts

The opening of the file, as written. The whole thing — 512 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Audit Deployed Package

Audit the package the way the reviewer audits it: as an installed artifact in a real org, not as source on disk. Source reading cannot see what install-time behavior does to the grants a subscriber actually receives, whether the post-install handler grants more than the packaged permission sets define, whether the data-flow engine flags CRUD/FLS on the compiled package, whether the callouts resolve against org-entered secrets, or whether uninstall leaves the org clean — and every one of those is a thing the Salesforce reviewer installs your package to check. This skill runs that pass and folds its findings into the same ledger the source audit feeds (.security-review/audit-ledger.json), under the deployed-package dimension, with evidence under .security-review/evidence/deployed-package/.

Honesty boundary, restated up front (CONVENTIONS §2): this is dynamic verification in a throwaway org — a complementary lens to static source reading, not a stronger one: it previews the reviewer's own install-into-their-org test by really installing and really querying the package, which source on disk cannot do (source cannot verify install-time behavior). But it is still not Salesforce's own penetration test. Salesforce's Product Security team installs your package and tests the surface regardless of anything this pass finds. Nothing here ever earns "will pass." The strongest thing it can say is "no known blockers in what this org let us verify."

When to use

  • AFTER /sf-security-review-toolkit:install-and-verify-package has installed a released version into a fresh org and its verification battery is green — that skill solves the hard part (headless install past every SF quirk, the credential config, the permission-chain battery). This skill is the thin security pass on top of an org that is already standing.
  • As the opt-in deep-audit power-up offered by /sf-security-review-toolkit:security-review-journey when sf is authed (or the operator opts to auth) — the journey spins a clean org, installs, runs this, then tears down. Never the always-on core.
  • After a package upgrade, re-run it: the install-time grant drop and the uninstall residue both change shape across versions.

Read the full file on GitHub · 512 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago First seen · 512 lines · 153 tokens per session scan A 1953d4266a4f

Subscribe to this mod's changes

audit-deployed-package is a skill published in the GitHub repository runverdict/sf-security-review-toolkit (2 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 153 tokens to every session and 8,486 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.