Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/PackmindHub/packmindnpx agentmods add skills/packmindhub/packmind/michel-run-local-dev-stackWrote 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/packmindhub/packmind/michel-run-local-dev-stack)<a href="https://agentmods.dev/skills/packmindhub/packmind/michel-run-local-dev-stack"><img src="https://agentmods.dev/badge/skills/packmindhub/packmind/michel-run-local-dev-stack/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/packmindhub/packmind/michel-run-local-dev-stack"><img src="https://agentmods.dev/badge/skills/packmindhub/packmind/michel-run-local-dev-stack.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 3 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 Data Exfiltration · line 102 Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
- medium Prompt Injection · line 205 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Data Exfiltration · line 222 Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.Fix: Verify the destination URL is trusted and necessary. Remove or replace with documented APIs. Ensure no secrets, tokens, or PII are transmitted.
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 | $0.00275 | $0.04996 |
| Opus 5.5 | $0.00110 | $0.01998 |
| Sonnet 5.5 | $0.00055 | $0.00999 |
| Haiku 4.5 | $0.00028 | $0.00500 |
Grade A, and why
michel-run-local-dev-stack scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
### What is reachable from the host — read this before you `curl` anything How it starts
The opening of the file, as written. The whole thing — 235 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Run the Packmind local dev stack
One reliable way to bring Packmind up locally, confirm it's serving, and take it back down. Other skills (michel-ui-demo-recorder, michel-cli-demo-recorder) and the Michel worker prompt all defer here instead of carrying their own copy — so the lifecycle stays correct in one place.
What the stack is
Defined by docker-compose.yml at the repo root. Every service runs from the base node:24.18.0-alpine3.23 image with the repo bind-mounted at /packmind — there is no app image to build. Code runs via nx serve/nx dev with polling file-watchers, so source edits hot-reload; you almost never pass --build.
What is reachable from the host — read this before you curl anything
Only these ports are published to the host: the frontend (4200 in OSS, 4201 in proprietary — see below), 443 (nginx), 5432 (postgres), 6379 (redis), 2345 (pgAdmin). The backend and mcp-server containers have no ports: mapping — their ports (3000 and 3001) exist only inside the compose network. curl localhost:3000 always fails from the host. This is the #1 runtime trap — do not fall into it.
Resolve the frontend host port — never hardcode 4200
The host-published frontend port differs by edition: 4200 for OSS, 4201 for proprietary (docker-compose.yml maps 4201:4200 there so a proprietary stack can run beside an OSS one without a clash). The container-internal port is always 4200 — so the Vite proxy, healthchecks, and e2e (frontend:4200) inside the compose network never change. Only the host port does. Ask compose for the real mapping instead of assuming; this is correct for either edition:
PM_WEB="$(docker compose port frontend 4200 | sed 's#.*:##')" # → 4200 (oss) or 4201 (proprietary)
Every host-side curl/URL below uses localhost:$PM_WEB. Resolve PM_WEB once after up -d (the container must exist for port to report the mapping) and reuse it.
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 · 235 lines · 275 tokens per session scan A 79720a5718c7
michel-run-local-dev-stack is a skill published in the GitHub repository PackmindHub/packmind (319 stars, last pushed yesterday), licensed Apache-2.0. It adds 275 tokens to every session and 4,996 once invoked, about $0.0011 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
harness
To build an AI harness: run, observe, validate, automate repeated work faster — CLI/MCP actions, devcontainers, skills, subagents, hooks, pipelines, automations.
ce-compound
Document a recently solved problem or durable project vocabulary in docs/solutions/ or CONCEPTS.md. Use when capturing a learning after work.
loadout-import-workflow
Import another repo's command/skill suite into loadout as a workflow. Use when the user points at an agent framework, plugin, or command pack (e.g. a GitHub repo, a .claude/commands/ directory, a skills pack) and wants its process turned into a loadout [[workflows]] entry — "import this workflow", "turn this plugin…
ce-commit-push-pr
Commit, push, and open a PR. Use when asked to ship/open a PR, or for PR-description-only flows like writing, rewriting, or describing a PR body.
sdd
Execute the Liatrio Spec-Driven Development (SDD) workflow when explicitly invoked by the user. NOTE: this skill is NOT intended to be dynamically loaded or automatically triggered; it should only ever be explicitly called by the user.
loadout-migrate
Migrate existing global AI agent instructions into loadout. Use when adopting loadout, importing a CLAUDE.md / AGENTS.md / global agent rules into loadout, setting up /.config/loadout/config.toml, or turning prose agent instructions into reusable fragments and stack-targeted profiles.