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 TalonT-Org/AutoSkillit --skill audit-feature-gatesgit clone --depth 1 https://github.com/TalonT-Org/AutoSkillitWrote 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/talont-org/autoskillit/audit-feature-gates)<a href="https://agentmods.dev/skills/talont-org/autoskillit/audit-feature-gates"><img src="https://agentmods.dev/badge/skills/talont-org/autoskillit/audit-feature-gates/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/talont-org/autoskillit/audit-feature-gates"><img src="https://agentmods.dev/badge/skills/talont-org/autoskillit/audit-feature-gates.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 | $0.00037 | $0.02745 |
| Opus 5 | $0.00018 | $0.01373 |
| Sonnet 5 | $0.00007 | $0.00549 |
| Haiku 4.5 | $0.00004 | $0.00275 |
Grade A, and why
audit-feature-gates 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.
How it starts
The opening of the file, as written. The whole thing — 278 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Audit Feature Gates Skill
Audits feature flag isolation across 6 dimensions: config projection, import chain integrity, runtime gate consistency, tool/skill tag completeness, boundary coupling, and test marker coverage. Detects where disabled features leak through import chains, runtime bypasses, UI surfaces, and ungated callables.
When to Use
- User says "audit feature gates", "check feature isolation", "feature flag audit"
- After adding a new feature to
FEATURE_REGISTRYto verify isolation is complete - As part of the
full-audit.yaml4th parallel chain
Arguments
No arguments required. Reads FEATURE_REGISTRY from src/autoskillit/core/_type_constants.py
to enumerate features.
Critical Constraints
NEVER:
- Modify any source code files
- Create files outside
{{AUTOSKILLIT_TEMP}}/audit-feature-gates/ - Issue subagent Task calls sequentially — ALL 6 must be in a single parallel message
- Write output files before synthesizing ALL subagent results
- Run subagents in the background (
run_in_background: trueis prohibited)
ALWAYS:
- Use
model: "sonnet"when spawning all subagents via the Task tool - Issue all 6 Task calls in a single message to maximize parallelism
- Subagents must NOT create their own files — they return findings in response text only
Severity Semantics
- BLOCK — active leakage that would execute feature code when the feature is disabled
- WARN — isolation gap that may cause confusion or is one step from leakage
- INFO — coupling assessment, informational only
Workflow
Step 0: Pre-flight — Enumerate Features
Read src/autoskillit/core/_type_constants.py. Extract all entries from FEATURE_REGISTRY exhaustively — do not assume a fixed list.
For each feature, note: name, lifecycle, import_package, tool_tags, skill_categories,
default_enabled.
Step 1: Launch 6 Parallel Subagents (SINGLE MESSAGE)
Issue ALL 6 Task calls in a single message.
D1 — Config Projection (subagent):
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.
- 9d ago First seen · 278 lines · 37 tokens per session scan A 21e6cfe465ab
audit-feature-gates is a skill published in the GitHub repository TalonT-Org/AutoSkillit (5 stars, last pushed today), licensed MIT. It adds 37 tokens to every session and 2,745 once invoked, about $0.0002 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.
Other skills, from other repositories
roundtrip-radar
Per-journey code audit tracing data through complete user flows for bugs, data safety, performance, and round-trip completeness. Discovers workflows, audits each end-to-end, rolls up cross-cutting issues, and supports natural-language flow tracing. Triggers: "roundtrip audit", "trace user journey", "/roundtrip-radar".
rust-intel
Hard rules for writing Rust in code that already compiles and passes tests but is silently broken, slow, or semver-fragile. Load this BEFORE writing any Rust code. Targets bugs that survive rustc, clippy, and cargo test but fail in production or rot the codebase. Covers async, unsafe, FFI, concurrency, crypto…
rust-intel
Hard rules for writing Rust in code that already compiles and passes tests but is silently broken, slow, or semver-fragile. Load this BEFORE writing any Rust code. Targets bugs that survive rustc, clippy, and cargo test but fail in production or rot the codebase. Covers async, unsafe, FFI, concurrency, crypto…
systematic-debugging
4-phase root cause debugging: understand bugs before fixing.
langsmith-observability
LLM observability platform for tracing, evaluation, and monitoring. Use when debugging LLM applications, evaluating model outputs against datasets, monitoring production systems, or building systematic testing pipelines for AI applications.
experimental-code-coverage-local-debugger
Runs code coverage locally via Universal Test Runner (UTR) or helper scripts, mimicking LUCI trybots. Activate when CQ tryjobs fail or underreport coverage, to test local GN/recipe repairs before uploading, or to debug hermetic crashes.