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 kreek/consult --skill debugginggit clone --depth 1 https://github.com/kreek/consultWrote 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/kreek/consult/debugging)<a href="https://agentmods.dev/skills/kreek/consult/debugging"><img src="https://agentmods.dev/badge/skills/kreek/consult/debugging/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/kreek/consult/debugging"><img src="https://agentmods.dev/badge/skills/kreek/consult/debugging.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 | $0.00028 | $0.00713 |
| Opus 5.5 | $0.00011 | $0.00285 |
| Sonnet 5.5 | $0.00006 | $0.00143 |
| Haiku 4.5 | $0.00003 | $0.00071 |
Grade A, and why
debugging 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 16d 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 — 63 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Debugging
Iron Law
NO FIX WITHOUT ROOT-CAUSE EVIDENCE.
When to Use
- Defects, flakes, timing-sensitive bugs, regressions, production incidents, unclear failures, or stuck debugging sessions.
When NOT to Use
- Planned refactors with no failing behavior; use
refactoring. - Slowness as the symptom; use
performance. - Bisect and reflog mechanics; pair with
git-workflow.
Rules
- Reproduce before fixing, or record why reproduction is not yet possible. Capture the exact symptom: command, input, output, stack trace, timing, environment.
- One hypothesis at a time. Predict what else must be true, run the smallest experiment that confirms or refutes it, and change one variable per experiment. After the third experiment, keep a short debug log.
- Before editing, state the failure model: likely cause, evidence for it, and the observation that would disprove it. A fix you cannot explain is not a fix.
- The root cause is named in one sentence, explains every observed symptom, and names the evidence that ruled out the main alternatives.
- The fix is one atomic change aimed at that cause, with a regression test or operational guard that fails before and passes after.
- A flake is a bug. Identify whether the test, the code, or the environment failed; never retry it away.
- Timing-sensitive bugs are verified with non-invasive observation or replay, not by adding sleeps.
- Incidents produce blameless learning with owned, dated follow-ups. "Human error" is never the root cause.
- Unproven root-cause or fix claims are reported as unproven.
Tripwires
| Trigger | Do this instead | False alarm |
|---|---|---|
| "Probably X, let me try fixing" | Gather evidence that confirms X before editing. | The user asked for a speculative explanation, not a fix. |
| "No time to reproduce" | Create the smallest reproduction or state why it is blocked. | Production-only incident where logs and traces are the reproduction. |
| "One more guess and it'll work" | Stop editing. Collect a new observation that changes the model. | A syntax or wiring typo visible in the failing output. |
| "Fixed it locally, ship it" | Name the root cause and run the regression guard. | The local run is the requested diagnostic, not a completion claim. |
| "Flake, just retry" | Treat it as a bug and find which of test, code, or environment failed. | Infrastructure outage already confirmed outside the code under review. |
| "Probably a race condition" | Show interleaving, shared state, or timing evidence before touching concurrency code. | A failing sanitizer or trace already demonstrates the race. |
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 16d ago Changed · +8 tokens per session e79baed1f108
- 1mo ago Changed · -19 lines a14202f0cf76
- 1mo ago First seen · 82 lines · 20 tokens per session scan A e7fb12f9d4c8
debugging is a skill published in the GitHub repository kreek/consult (1 stars, last pushed 17d ago), licensed MIT. It adds 28 tokens to every session and 713 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 2026-08-31.
Other skills, from other repositories
diagnose-project
Diagnose .NET SDK, restore, build, test, package, target framework, F# compile-order, C# analyzer, and project-reference failures with concrete next checks.
bug-fix
Use when a bug, regression, or unexpected behaviour is reported and the user wants it fixed end to end — "why is X broken", "this stopped working after Y", "fix this crash". Also use when a symptom is known but its cause is not. Not for building new behaviour, and not for a change whose cause is already proven.
bug-investigate
Use when someone wants to know WHY something is broken and has not asked for it to be fixed — "why is this happening", "what's causing this error", "find out what's wrong", "diagnose this before we decide". Also use before committing to a fix, when the cause is unknown and the decision depends on it. Not for fixing …
xcode-build-run-workflow
Guide build, run, preview-adjacent, workspace-inspection, diagnostics, toolchain, and guarded mutation work in existing Xcode-managed projects and workspaces. Use when Xcode-aware execution is needed and the task is primarily about build, run, or project-integrity work rather than testing.
xcode-debugger-mcp-workflow
Diagnose a running Apple app through Xcode’s active LLDB debugging session and assess the experimental Xcode 27 lldb-mcp executable. Use when setting breakpoints, inspecting frames or variables, issuing LLDB commands, or distinguishing a usable Xcode debugger session from an unavailable standalone debugger MCP server.
safari-mcp-workflow
Inspect a scoped live site in Safari through Safari Technology Preview's local MCP server. Use for Safari-specific DOM, console, network, screenshot, accessibility, performance, interaction, responsive, or print-media evidence.