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 agentmods add skills/angad-kandhari/deliberate/debugnpx skills add angad-kandhari/deliberate --skill debuggit clone --depth 1 https://github.com/angad-kandhari/deliberateWrote 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/angad-kandhari/deliberate/debug)<a href="https://agentmods.dev/skills/angad-kandhari/deliberate/debug"><img src="https://agentmods.dev/badge/skills/angad-kandhari/deliberate/debug.svg" alt="Measured on agentmods" 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 | $0.00066 | $0.01329 |
| Opus 5 | $0.00033 | $0.00665 |
| Sonnet 5 | $0.00013 | $0.00266 |
| Haiku 4.5 | $0.00007 | $0.00133 |
Grade A, and why
debug 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 4d 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 — 141 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Debug
Debugging skill for LLM coding agents. Load this when the task is "fix the bug," "it's broken," "tests are failing," or any request that starts from a symptom rather than a feature.
Where deliberate keeps code honest, this skill keeps diagnosis honest. Counters the failure mode of trying random fixes until symptoms go away.
1. Reproduce Before You Reason
No repro, no fix.
- Get the bug to happen on demand before touching code. A failing test is ideal; a manual repro with exact steps is the floor.
- If you can't reproduce, that's the first bug to fix. Don't guess at root cause from a log snippet.
- Write the repro down. It's the acceptance test for your fix.
If the user hands you a flaky or intermittent bug, say so:
"I can trigger this maybe 1 in 5 times. Before fixing, I want to find conditions that make it reliable - otherwise I can't tell if a fix worked."
Test: Can I make this bug happen whenever I want?
2. Symptom, Cause, Fix - Keep Them Separate
The symptom is not the bug.
State all three out loud before editing:
- Symptom: what the user sees (error message, wrong output, crash).
- Cause: why it happens (the actual defect in logic, state, or assumption).
- Fix: the minimum change that addresses the cause.
If you can name the fix but not the cause, you're pattern-matching, not debugging. Stop.
Test: Could I explain to a new engineer why this bug existed, not just what I changed?
3. Bisect, Don't Guess
Narrow the search space systematically.
When the cause isn't obvious:
- Git bisect when a behavior regressed. It's faster than reading diffs.
- Binary search the code path: add a log or assertion halfway through, see which side the bug is on, repeat.
- Minimize the repro until only the bug-relevant code remains.
- Diff a working case against a broken case - the delta is the bug, almost always.
Random print statements across the codebase is not bisection. It's noise.
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.
- 4d ago First seen · 141 lines · 66 tokens per session scan A 441bc74b5664
debug is a skill published in the GitHub repository angad-kandhari/deliberate (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 66 tokens to every session and 1,329 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.
Other skills, from other repositories
skill-monitor
Analyze skill effectiveness across sessions. Computes per-skill metrics (action rate, friction, outcomes), identifies degrading skills, and generates improvement recommendations. Requires session-scan data in metrics.jsonl.
audit
Project health audit and health check — architecture, performance, tests, dependencies, code quality. Use when assessing overall project health, before releases, or after refactors.
promote
Generate X/Twitter release promotion posts with ASCII tables and CodeSnap rendering. Use when writing release posts, promotion tweets, plugin announcements, or preparing social media content for new versions.
release
CONTRIBUTOR TOOL - Cut a plugin release: bump plugin.json version, finalize CHANGELOG, update README if needed, gate on make ci, commit, tag vX.Y.Z, and create the GitHub release. Use when shipping a new plugin version. NOT distributed.
session-deep-dive
Deep qualitative analysis of high-signal sessions. Spawns subagents with v2 template, synthesizes patterns, compares against known findings. Use after /session-scan.
catchup
Summarize and review what changed while you were away. Use after a weekend, vacation, or flight to check missed PRs, git commits, Linear tickets, and meetings — one prioritized brief, not a firehose.