Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add tarlepp/symfony-flex-backend --skill add-security-votergit clone --depth 1 https://github.com/tarlepp/symfony-flex-backendWrote 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.
[](https://agentmods.dev/skills/tarlepp/symfony-flex-backend/add-security-voter)<a href="https://agentmods.dev/skills/tarlepp/symfony-flex-backend/add-security-voter"><img src="https://agentmods.dev/badge/skills/tarlepp/symfony-flex-backend/add-security-voter/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.
<a href="https://agentmods.dev/skills/tarlepp/symfony-flex-backend/add-security-voter"><img src="https://agentmods.dev/badge/skills/tarlepp/symfony-flex-backend/add-security-voter.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00030 | $0.01449 |
| Opus 5.5 | $0.00012 | $0.00580 |
| Sonnet 5.5 | $0.00006 | $0.00290 |
| Haiku 4.5 | $0.00003 | $0.00145 |
Grade A, and why
add-security-voter 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
What is this?
Use this skill when adding or changing a Symfony security voter, its supported
attributes or subjects, or the authorization checks that invoke it. For changes
to credentials, authentication flows, firewalls, or broader role policy, also
follow .github/skills/implement-api-authentication/SKILL.md.
Follow the shared skill guidance, including
.github/copilot-instructions.md and doc/AI_RULES.md. Read doc/SECURITY.md
for the repository security policy.
Table of Contents ᐞ
Workflow ᐞ
- Read
composer.jsonfor supported PHP and Symfony Security versions andsymfony.lockfor recipes. Inspectconfig/packages/security.yaml,config/services.yaml,src/Security/Voter/IsUserHimselfVoter.php, its tests, and every affected authorization call site before choosing a pattern. Trace how the subject is resolved and which authenticated identity reaches the voter; do not infer either from the route or entity class alone. - Establish the permission attribute, supported subject type, allowed and denied callers, ownership rules, and any role-based override. Include anonymous callers, other owners, JWT users, and API-key callers where relevant. Ask before making an unspecified policy decision or changing an access boundary. Extend an existing voter when it owns the same policy; keep unrelated permissions separate.
- Place voters in
src/Security/Voter/and extend Symfony'sVoter, following local typing, generics, and#[Override]conventions. Match the installed method signatures, including the optionalVoteargument where applicable. Implementsupports()narrowly for the exact attributes and subject types. Unsupported requests must abstain, not grant or deny unrelated permissions; denial of a supported request belongs invoteOnAttribute(). Account for subjectless checks only when the intended permission requires them. - Read the caller from the supplied
TokenInterface. Human users are represented bySecurityUser, not the DoctrineUserentity; API keys useApiKeyUser. For ownership checks, compare stable entity identifiers asIsUserHimselfVoterdoes, not object identity or display names. Do not treat an API-key identifier or credential as a human user's UUID. Deny supported requests for missing or inappropriate identities unless the policy explicitly permits them. Never log credentials or expose them in denial messages or tests. - Keep authorization decisions focused and free of mutations. Inject existing
resources or services when domain information is needed rather than putting
persistence or business operations in the voter. If the policy depends on
roles, preserve configured inheritance using the existing security decision
or role helpers; do not replace hierarchy-aware checks with a raw
getRoles()membership test. When delegating another decision from a voter, use the supplied token withAccessDecisionManagerInterfaceand avoid voting recursively on the same attribute and subject. - Preserve the configured
unanimousaccess-decision strategy and inspect overlapping voters: one supported denial can block an otherwise granted permission. Do not change the strategy or all-abstain behavior to make a new check pass. Keep intentional alternatives together at the correct decision boundary. For example,UserRolesControllerandUserGroupsControllerexpress self-or-root access with anIsGrantedexpression; adding a separate ownership requirement alongside a role check is not equivalent to that alternative. - Wire the permission at every relevant caller using the established
#[IsGranted]or authorization-service pattern. Pass the actual subject; for attribute-based controllers, match the subject name to the resolved controller argument. Preserve class-level, method-level, trait-based, and firewall guards unless the requested policy explicitly changes them. Role-only checks do not enforce ownership, and merely creating a voter does not protect an endpoint. Preserve existing response status and error handling rather than swallowing access-denied exceptions. - Rely on existing
App\service discovery andautoconfigure: truefor standard voter registration. Do not add redundantsecurity.votertags or manually instantiate a voter at application call sites. If service configuration is exceptional, inspect its registration before changing it. Keep request-specific identities and subjects out of shared service state. - Follow
.github/skills/write-tests/SKILL.md. Usetests/Integration/Security/Voter/IsUserHimselfVoterTest.phpas a reference for exercising the publicvote()boundary and asserting grant, deny, and abstain results, not protected methods. Cover unsupported attributes and subjects, anonymous or inappropriate identities, matching and different owners, and relevant role overrides. Add container-backed decision-manager coverage when registration, voter overlap, or role inheritance matters. For endpoint checks, followtests/E2E/Controller/v1/User/UserRolesControllerTest.php: exercise the real request with allowed and denied callers and assert response status and protected output or absence of write side effects. Test JWT and API-key paths when shared policy affects both. Direct voter tests alone do not prove registration, correct subject binding, or endpoint protection. - Update relevant security documentation and OpenAPI access/error metadata when the observable authorization contract changes. Keep policy details close to existing documentation rather than creating parallel guidance.
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.
- yesterday First seen · 123 lines · 30 tokens per session scan A fa97e3acc6c1
add-security-voter is a skill published in the GitHub repository tarlepp/symfony-flex-backend (288 stars, last pushed today), licensed MIT. It adds 30 tokens to every session and 1,449 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-10-21.
Other skills, from other repositories
api-integration-architect
Design, implement, debug, and optimize API integrations with expert-level patterns for REST, GraphQL, webhooks, and authentication flows.
horse-database-pooling
Guide for setting up thread-safe database connection pooling (FireDAC / UniDAC) in multithreaded Horse applications.
horse-request-response
Guide to interacting with THorseRequest (body, query, params, headers) and THorseResponse (Send, Status, ContentType).
horse-dependency-injection
Guide for managing request-scoped contextual services and IoC (dependency injection) in Delphi and Lazarus.
horse-mvc-architecture
Guide for structuring corporate Horse applications using Clean MVC (Model-View-Controller) principles and decoupling HTTP layers from business logic.
horse-app-structure
Guide for setting up Horse applications, bootstrap program (.dpr), basic console initialization, and registering modules.