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 angad-kandhari/deliberate --skill incidentgit 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/incident)<a href="https://agentmods.dev/skills/angad-kandhari/deliberate/incident"><img src="https://agentmods.dev/badge/skills/angad-kandhari/deliberate/incident/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/angad-kandhari/deliberate/incident"><img src="https://agentmods.dev/badge/skills/angad-kandhari/deliberate/incident.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 | $0.00065 | $0.01595 |
| Opus 5 | $0.00032 | $0.00797 |
| Sonnet 5 | $0.00013 | $0.00319 |
| Haiku 4.5 | $0.00006 | $0.00160 |
Grade A, and why
incident 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 10d 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 — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Incident
Incident-response skill for LLM coding agents. Load this when production is broken, something is on fire, a paging alert is active, or the user is debugging under time pressure.
Where deliberate keeps normal work honest, this skill keeps crisis work honest. Counters the two dominant failure modes under pressure: making it worse by acting too fast, and freezing by analyzing too long.
1. Stabilize First, Diagnose Second
Stop the bleeding before you look for the wound.
When production is broken, the first goal is "make it stop being broken for users." Root cause is the second goal.
Stabilization moves, in rough order of preference:
- Revert. Roll back the change that started it. If the timing fits, just do it.
- Feature flag off. Disable the feature path without a deploy.
- Failover. Shift traffic to a healthy region, replica, or fallback.
- Throttle / shed. Reduce load to something the system can handle.
- Restart. Kick the process. Crude but fast.
- Last resort: ad-hoc fix. Highest risk of making it worse.
"Let me figure out why" when users are down is the wrong instinct. Stabilize, then figure out.
Test: If I could stop user impact right now with one action, what would it be?
2. Communicate While You Work
Silence during an incident costs more than incomplete information.
- Post an early ack in the incident channel: "I see it, I'm on it, will update in 10 min."
- Update every 10-15 minutes even if nothing changed. "Still investigating, ruled out X and Y" is useful.
- Name what you know, what you suspect, and what you don't know. Keep them separate.
- When handing off, hand off the state of the investigation, not just the symptoms.
Stakeholders fill silence with worst-case assumptions. Don't let them.
Test: Does anyone who joins the incident channel right now know what's going on without asking?
3. Don't Lose Evidence
The moment you touch production, you may destroy the thing that would have told you the cause.
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.
- 10d ago First seen · 159 lines · 65 tokens per session scan A 2bbc5916d786
incident is a skill published in the GitHub repository angad-kandhari/deliberate (2 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 65 tokens to every session and 1,595 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
x-bug2rag
A knowledge-capture tool that turns reusable bug explanations into a local RAG collection, meaning a searchable store of text that an agent can retrieve later. It records the trigger, incorrect implementation, correct implementation, and observable difference.
x-cr
A software-correctness investigation skill for finding why code behaves differently from what was expected. It uses evidence from code paths, specifications, tests, logs, and changes to assess possible causes.
x-fix
A bug-fixing skill with separate paths for directly reported bugs, correctness-review findings, and failed verification or quality-gate reports. It records completed fixes and their validation in project reports.
x-audit-perf
A manually triggered review of a whole software project’s performance. It looks for code and system patterns that may make programs slower or use more resources.
issue-debugging
Systematic methodology for issue debugging including root cause analysis, impact mapping, tiered validation plans, and confidence assessment. Use when analyzing bugs, fixing issues, or validating fixes.
fix-bug
Run the Fix Validation pipeline to investigate, fix, and validate a bug. Ensures deterministic pipeline execution with IssueAnalyzer, FixWriter, TestWriter (conditional), TestAudit (conditional), and FixValidator stages.