Borrowing it
Nothing to install: this file belongs to GabrielOnDelphi/Claude-Tools-for-Delphi. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/GabrielOnDelphi/Claude-Tools-for-Delphi/main/.claude/skills/light-task-DerailmentCheck/SKILL.mdgit clone --depth 1 https://github.com/GabrielOnDelphi/Claude-Tools-for-DelphiWrote 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/gabrielondelphi/claude-tools-for-delphi/light-task-derailmentcheck)<a href="https://agentmods.dev/skills/gabrielondelphi/claude-tools-for-delphi/light-task-derailmentcheck"><img src="https://agentmods.dev/badge/skills/gabrielondelphi/claude-tools-for-delphi/light-task-derailmentcheck/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/gabrielondelphi/claude-tools-for-delphi/light-task-derailmentcheck"><img src="https://agentmods.dev/badge/skills/gabrielondelphi/claude-tools-for-delphi/light-task-derailmentcheck.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 | $0.00188 | $0.01158 |
| Opus 5 | $0.00094 | $0.00579 |
| Sonnet 5 | $0.00038 | $0.00232 |
| Haiku 4.5 | $0.00019 | $0.00116 |
Grade A, and why
light-task-DerailmentCheck 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.
How it starts
The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/light-task-DerailmentCheck — pause and verify conclusions
A subagent's report, a plausible-looking fix, and a confident root cause all share the same failure mode: they feel verified without being verified. This protocol forces the distinction before more work gets built on top of an unchecked assumption. It costs a few minutes; a wrong conclusion discovered three steps later costs the whole afternoon.
Run the Full version whenever this skill is invoked directly. Use the Short version only as a quick inline nudge when you are checkpointing yourself mid-task without stopping to run the full audit as a separate step.
Full version
STOP. Do not write, edit, or run anything until this audit is done.
- Original goal. Re-read the first message of this task. One sentence: what was asked for? Is the CURRENT activity still serving it, or did the scope drift?
- Load-bearing conclusions. List every conclusion drawn so far that the current plan depends on. One line each.
- Classify each: VERIFIED (direct proof seen — cite file:line / command output / doc URL), INFERRED (plausible deduction, never directly checked), ASSUMED (no evidence, it just felt right). A subagent's report counts as INFERRED until independently re-checked.
- Attack. For each INFERRED / ASSUMED item, actively try to DISPROVE it — read the real code, run the check, search the docs. Argue against it like a hostile reviewer. Promote to VERIFIED or mark it WRONG. Nothing the next step depends on may stay unverified.
- Earliest wrong turn. Find the FIRST item that is now WRONG. Everything built on top of it is suspect — name the work that must be discarded.
- Verdict:
- ON TRACK — every load-bearing conclusion is VERIFIED. Continue.
- DERAILED — state the wrong turn in one sentence, go back to the last verified point, and show the corrected plan before touching anything else.
Do not defend prior work. A wrong conclusion found now is cheap; a wrong conclusion kept is expensive.
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.
- 2d ago Changed · +3 lines · -9 tokens per session 8271cdbfeffa
- 5d ago Changed · +2 lines 74c150064fc2
- 10d ago First seen · 50 lines · 197 tokens per session scan A 190066d103e2
light-task-DerailmentCheck is a skill published in the GitHub repository GabrielOnDelphi/Claude-Tools-for-Delphi (17 stars, last pushed 2d ago), licensed MPL-2.0. It adds 188 tokens to every session and 1,158 once invoked, about $0.0009 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 2026-08-30.
Other skills, from other repositories
debug-optimize-lcp
Guides debugging and optimizing Largest Contentful Paint (LCP) using Chrome DevTools MCP tools. Use this skill whenever the user asks about LCP performance, slow page loads, Core Web Vitals optimization, or wants to understand why their page's main content takes too long to appear. Also use when the user mentions…
systematic-debugging
Use when debugging a failing test, build error, or runtime issue that isn't immediately obvious. Guides a 4-phase root cause analysis instead of random fix attempts.
diagnose
Trace from a reproduced symptom to the source code that causes it. Pin the specific file and approximate line, rate confidence in the cause and clarity of the fix independently, and always propose a concrete fix.
repro-admin
Reproduce an EmDash admin UI bug. Attach a container, start the demo dev server, drive the admin with agent-browser using the dev-bypass session, and capture the reproduction as screenshots plus a replayable transcript.
log-error-digest
Analyze log files to troubleshoot errors, identify peak error periods, and produce error clustering, frequency statistics, and time distribution reports. Supports JSON, syslog, and Nginx formats with automatic detection. Use when a user uploads a .log file and asks to analyze errors, find patterns, debug issues, or…
byted-util-volcengine-detect-retry
An orchestration workflow for Volcengine Cloud Detect, a service that checks websites or network endpoints from test locations.