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 Kevin-Liu-01/Agent-Machines --skill postmortemgit clone --depth 1 https://github.com/Kevin-Liu-01/Agent-MachinesWrote 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/kevin-liu-01/agent-machines/postmortem)<a href="https://agentmods.dev/skills/kevin-liu-01/agent-machines/postmortem"><img src="https://agentmods.dev/badge/skills/kevin-liu-01/agent-machines/postmortem/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/kevin-liu-01/agent-machines/postmortem"><img src="https://agentmods.dev/badge/skills/kevin-liu-01/agent-machines/postmortem.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 34 Skill grants unrestricted tool access without appropriate constraints. An agent with unfettered tool access can perform arbitrary actions including file modification, network requests, and code execution.Fix: Restrict tool access to only the tools required for the skill's stated purpose. Use an explicit allowlist rather than granting blanket access.
- medium Excessive Agency · line 115 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00057 | $0.02111 |
| Opus 5.5 | $0.00023 | $0.00844 |
| Sonnet 5.5 | $0.00011 | $0.00422 |
| Haiku 4.5 | $0.00006 | $0.00211 |
Grade A, and why
postmortem 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 1mo 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 — 245 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Postmortem
Structured incident response. Investigate first, write second, verify third.
When to use
- Production outage or degradation
- Financial loss (billing bugs, fraud, credit abuse)
- Security incident (unauthorized access, data leak, bot abuse)
- Silent failure discovered after the fact (broken integration, dead code path)
- Any event where the team needs a shared written record of what happened
Philosophy
A postmortem is a forensic document, not a blame report. It answers:
- What happened? (timeline, evidence, impact)
- Why did it happen? (root causes, not symptoms)
- How do we know the fix works? (verification queries, not vibes)
- How do we prevent recurrence? (structural fixes, not patches)
Every claim must be backed by data. "We think X" is not acceptable. "Query Y returned Z, which proves X" is acceptable.
Workflow
Phase 1: Investigate
Gather evidence before forming theories. Use every tool available.
Database queries (/db skill):
- Query production data to quantify impact (how many users, how much money, what time range)
- Cross-reference tables to find patterns (which accounts, what behavior, what correlation)
- Always use
--readonlyand--env prod(only when the user confirms production) - Save the queries in the postmortem so anyone can re-run them
Git history (git log, git blame):
- Build a commit timeline for the affected code paths
- Identify who changed what and when (for context, not blame)
- Find the commit that introduced the bug vs. commits that tried to fix it
- Use
git log --format="%h %ad %an %s" --date=shortfor clean timelines
External service dashboards (Stripe, Clerk, Vercel, AWS):
- Use MCP tools when available (Stripe MCP, Supabase MCP, etc.)
- Use CLI tools (
gh,aws,stripe) when MCP is not available - Screenshot or copy dashboard data as evidence
Code reading:
- Trace the full code path from entry point to failure
- Read the actual production code, not what you think it does
- Check for race conditions, silent error swallowing, missing validation
- Verify claims against the code ("the threshold is 90" -- is it really?)
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.
- 1mo ago First seen · 245 lines · 57 tokens per session scan A b39e4ceb71c3
postmortem is a skill published in the GitHub repository Kevin-Liu-01/Agent-Machines (30 stars, last pushed 9d ago), licensed MIT. It adds 57 tokens to every session and 2,111 once invoked, about $0.0002 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 2026-09-03.
Other skills, from other repositories
build-test
Run the project's build / typecheck / lint / test commands and emit the build.passing + tests.passing signals devloop convergence reads.
systematic-bug-hunting
Use when actually hunting for exploitable bugs at the bench inside an already-authorized engagement — reading a target, running scanners/fuzzers, sitting on a pile of odd-but-inert observations, tempted to report something "probably exploitable", or about to conclude "nothing found". Applies to every target class…
opik-explain
Root-cause a specific Opik trace, or a pattern across traces, and return a grounded explanation. Uses the hosted Opik MCP when it is connected, and falls back to SDK scripting otherwise. Returns the root cause, the evidence spans as clickable Opik UI links, and one suggested next step. Use for "why did this trace…
ssh-incident
A diagnosis order for "the server is down", "the site is slow", "something broke". Use when investigating a problem on a remote server through the ssh tools, before changing anything.
node-inspect-debugger
Debug Node.js with node inspect, inspector attach, breakpoints, Chrome DevTools Protocol, heap snapshots, and CPU profiles.
dependency-cycle-audit
Use when module, package, or build dependencies may contain cycles, especially load-order failures, broad rebuilds, or a planned extraction: derive the real graph, compute strongly connected components, rank cycle clusters, choose the minimal feedback-edge cut, and add an acyclicity guard.