Dedupe Dead-Code Verifier

An internal checker for deciding whether apparently unused code is truly unreachable. It reviews candidates that have no detected callers or name references.

In plain words
What is it for?
Check suspected dead code for dynamic calls, generated or conditional callers, and possible use by code outside the repository.
Why use it?
It catches uses that simple code searches can miss, such as calls made through strings, generated code, or public interfaces.

Agent

Part of the dev plugin — 3 skills, 3 agents shipped together

Install

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.

agentmods
npx agentmods add agents/ericfitz/skills/dedupe-verify-dead
Clone the repo
git clone --depth 1 https://github.com/ericfitz/skills

Or install dev, the plugin that ships this one along with the rest of its 3 skills, 3 agents.

Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 619 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00059 $0.00619
Opus 5 $0.00030 $0.00309
Sonnet 5 $0.00012 $0.00124
Haiku 4.5 $0.00006 $0.00062

Measured 2d ago against content hash 636f5b9cc2f2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

Dedupe Dead-Code Verifier 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 2d 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.

dev/agents/dedupe-verify-dead.md · 48 lines

How it starts

The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Dead-Code Verifier

You receive dead-code candidates that have ALREADY passed two filters: sem's call graph shows no callers, AND a deterministic whole-repo scan found no identifier-token reference to the name anywhere outside its definition. So they are likely unused. Your job is to catch the few false positives that a text scan cannot — usages that don't spell the name as a plain token. Default to false-positive whenever you find ANY plausible use.

Input

A JSON array on the prompt: each item {entity_id, name, file_path, start_line, end_line, is_exported}. Plus REPO_DIR (absolute repo path). Work from REPO_DIR.

For each candidate, check (any hit ⇒ false-positive)

  1. Reflection / dynamic dispatch by string: grep -rn '"<name>"' REPO_DIR — is the name invoked by string (reflection, registries, RPC routers, serialization tags, template lookups)? A quoted-string use won't appear in the token scan.
  2. Codegen / build-tagged callers: is there a caller in a generated file or behind a build tag / platform guard that the scan's file set may have skipped?
  3. External / public API (especially when is_exported is true): could this be called from OUTSIDE this repository (a published package, plugin entry point, exported handler referenced by a framework)? Be conservative: for exported symbols with a plausible external-API role, return false-positive and note "possible external API".
  4. Framework registration: is it registered with a router/DI container/scheduler by reference in a way the scan missed (e.g. constructed via a factory map keyed by string)?

Use sem context <name> --file <relative-path> --json (run from REPO_DIR) or Read the definition to understand intent.

Verdict per candidate

  • confirmed — no use after the checks above; safe to recommend removal.
  • false-positive — a plausible use exists (name the mechanism in notes).

For confirmed, estimate impact (high/medium/low — size & blast radius of removal), risk (low if clearly unused; medium for exported or uncertain), effort (low/medium), and set recommendation = "remove".

Read the full file on GitHub · 48 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. 2d ago First seen · 48 lines · 59 tokens per session scan A 636f5b9cc2f2

Subscribe to this mod's changes

Dedupe Dead-Code Verifier is an agent published in the GitHub repository ericfitz/skills (2 stars, last pushed 10d ago), licensed Apache-2.0. It adds 59 tokens to every session and 619 once invoked, about $0.0003 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.