Skip to content
Back to skills

Pre Mortem

ASecurity

Launches a fresh pre-mortem agent on a settled discuss and returns the one-line status of its result. Use once the phase is in its premortem stage, before the plan.

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 11, 2026
ai-agents

Security analysis

A100/100

Scanned November 11, 2026

npx -y skills add danitrrga/forge --skill pre-mortem --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Pre Mortem?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Pre Mortem
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/danitrrga-pre-mortem/badge)](https://www.skillsdirectory.com/skills/danitrrga-pre-mortem)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: pre-mortem
description: Launches a fresh pre-mortem agent on a settled discuss and returns the one-line status of its result. Use once the phase is in its premortem stage, before the plan.
user-invocable: false
---
<!-- referent: pstack plugins/pstack/skills/thermos/SKILL.md @36885ea -->
<!-- Differs from the referent: one agent, not two, and in the foreground because the status needs its file; the agent reads its own brief, so the brief never enters this conversation; the model comes from the fwx configuration. -->

You run the pre-mortem of one phase and hand back its status. The session writes the stages and talks with the owner; you do neither, and you never open `premortem.md`, so it stays out of this conversation.

You are given the phase folder, the language of the conversation, and whether anyone can answer (`--unattended` when nobody can; then English).

1. Run `fwx plan brief <dir> premortem --head`. It prints `{agent, model, round}`. On exit 1 the phase is not ready (no `premortem` stage row, no `context.md`) and on exit 2 the folder or configuration cannot be used: return the message and stop.
2. Launch the `Agent` tool with that `agent` as `subagent_type` and that `model`, unchanged: the owner set the model for this size of work, and a default from elsewhere in the session does not replace it. The prompt is "Run `fwx plan brief <dir> premortem` and follow it." with `--unattended` added if it applies, then "Write in <language>." A launch returns at once: wait for the agent's completion notice before step 3.
3. Run `fwx plan status <dir> premortem`, with `--unattended` if it applies. On `failed`, launch the agent once more the same way and run the status again.
4. End your reply with the line step 3 printed, copied as it is and with nothing after it, because the session reads that line.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.