How do I install the Roast skill?

Run npx skillmds@latest add acedergren/roast in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Roast safe to use?

SkillMD's automated safety review verdict for this skill is PASS. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Roast?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Roast free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

acedergren (@acedergren) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Roast by acedergren — Agent Skill

Name: Roast
Author: acedergren

File contents

Agent Workflow Roast

Use this skill when the user asks for @roast, /roast, or a coaching report for recent agent work.

Workflow

If the user provides existing Agent Workflow Roast JSON, markdown, HTML, aggregate metrics, or redacted evidence and asks for interpretation, work from that provided material only. Do not read raw session JSONL, memory files, history files, or secrets unless the user explicitly asks for a fresh /roast run.

Capture the folder where the skill was triggered before changing directories. From the plugin root, run:
```
AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs
```

Pass through user options when supplied:

AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --days 30 --no-memory
AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --project oci-self-service-portal
AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --no-ai
AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --export markdown
AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --site

Share the generated report path. The HTML report is always named agent-workflow-roast.html and written to the folder where the skill was triggered, unless the user explicitly supplies an output directory. If the user asks for Sites presentation, run with --site and upload the resulting agent-workflow-roast-site.tgz archive to Sites.

Notes

The analyzer redacts obvious secrets before synthesis or rendering.
Treat precomputed metrics and redacted report content as enough for follow-up coaching; ask for a fresh run only when the user wants updated data.
Treat the playground as legacy prompt/design scaffolding. For presenting generated reports, use the --site output path and Sites deployment.
Missing history, session, or memory files should degrade into a sparse report instead of failing.
If qualitative synthesis is unavailable, use the deterministic report sections.

---name: roastdescription: Generate an Agent Workflow Roast report from recent local session history and optional memory context. Use when the user invokes @roast, /roast, or asks for workflow coaching with receipts.---# Agent Workflow RoastUse this skill when the user asks for `@roast`, `/roast`, or a coaching report for recent agent work.## WorkflowIf the user provides existing Agent Workflow Roast JSON, markdown, HTML, aggregate metrics, or redacted evidence and asks for interpretation, work from that provided material only. Do not read raw session JSONL, memory files, history files, or secrets unless the user explicitly asks for a fresh `/roast` run.1. Capture the folder where the skill was triggered before changing directories. From the plugin root, run: ```bash AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs ```2. Pass through user options when supplied: ```bash AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --days 30 --no-memory AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --project oci-self-service-portal AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --no-ai AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --export markdown AGENT_WORKFLOW_ROAST_OUTPUT_DIR="$TRIGGER_DIR" node scripts/agent-workflow-roast.mjs --site ```3. Share the generated report path. The HTML report is always named `agent-workflow-roast.html` and written to the folder where the skill was triggered, unless the user explicitly supplies an output directory. If the user asks for Sites presentation, run with `--site` and upload the resulting `agent-workflow-roast-site.tgz` archive to Sites.## Notes- The analyzer redacts obvious secrets before synthesis or rendering.- Treat precomputed metrics and redacted report content as enough for follow-up coaching; ask for a fresh run only when the user wants updated data.- Treat the playground as legacy prompt/design scaffolding. For presenting generated reports, use the `--site` output path and Sites deployment.- Missing history, session, or memory files should degrade into a sparse report instead of failing.- If qualitative synthesis is unavailable, use the deterministic report sections.

Frequently asked questions

Roast