Is steelman safe to install?

SaferSkills independently audited steelman (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.

CategoryWeightCategory scoreContribution

Securityprompt, exec, net, exfil, eval

35%

100

35.0 pts

Supply chainhash, typosquat, maintainer, lockfile

20%

100

20.0 pts

Maintenancestaleness, pinning, CI

15%

100

15.0 pts

TransparencySKILL.md, perms, README

15%

100

15.0 pts

Communityinstalls, verify, response

15%

100

15.0 pts

Findings & checks · 0 flagged

Securityscore 100 · 0 findings

✓—All security checks passedNo findings in this category for the latest scan.pass

Supply chainscore 100 · 0 findings

✓—All supply chain checks passedNo findings in this category for the latest scan.pass

Maintenancescore 100 · 0 findings

✓—All maintenance checks passedNo findings in this category for the latest scan.pass

Transparencyscore 100 · 0 findings

✓—All transparency checks passedNo findings in this category for the latest scan.pass

Communityscore 100 · 0 findings

✓—All community checks passedNo findings in this category for the latest scan.pass

The primary manifest — the file an agent reads to learn what this artifact does.

skills/steelman/SKILL.md · 1 file

skills/steelman/SKILL.md2.0 KB · Markdown

name: steelman

description: Build the strongest version of an opposing view or rejected option. Use when the user has dismissed an alternative and you want to test that dismissal honestly. Produces the best-faith argument for the rejected position, then a head-to-head comparison.

last_reviewed: 2026-05-09

steelman

The opposite of strawmanning. Build the strongest version of the view you're about to disagree with, then engage with that.

When to use

User has dismissed an alternative quickly ("Redux is dead", "we

shouldn't use Tailwind", "monorepos are over").

User is choosing between options and has only argued for one side.
After convince-me — once the user's case is on the table, build

the opposing case.

When NOT to use

The dismissed option is genuinely terrible and a steelman would be

dishonest. (Say so.)

The user wants a recommendation, not a debate.

Stance

Genuinely sympathetic to the position you're steelmanning. Don't

hedge.

Voice the position as a smart practitioner who's chosen it on

purpose. What problem are they solving that the user isn't?

After the steelman, do a head-to-head. Don't disappear into pure

advocacy.

Procedure

Restate the rejected option in its best-case form, including the

constraints under which it's the right choice.

List the 3 strongest reasons a thoughtful practitioner would pick

it.

List the strongest counter to the user's chosen option from the

steelmanned position's point of view.

Compare head-to-head: which constraints favour each side?
Recommend: under what conditions should the user revisit?

Example openers

"Here's the strongest case for the option you just rejected — built

in good faith:"

Output format

The steelman: <one paragraph>

Three reasons a thoughtful practitioner picks <rejected option>:
1. …
2. …
3. …

Counter to your chosen option from that view:
- …

Head-to-head:
- Constraint X favours: <option>
- Constraint Y favours: <option>

When to revisit: <conditions>

steelman— agent skill

Is steelman safe to install?

steelman

When to use

When NOT to use

Stance

Procedure

Example openers

Output format

Audit the pieces. Scan the whole. Decide.