AgentStack
SKILL verified MIT Self-run

Thoughts

skill-wave-engineering-claudecode-workflow-thoughts · by Wave-Engineering

The user is inviting a genuine alignment check on a proposal — stress-test it for real, then act on the result in ONE turn (aligned → proceed; concern → surface it)

No reviews yet
0 installs
4 views
0.0% view→install

Install

$ agentstack add skill-wave-engineering-claudecode-workflow-thoughts

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Thoughts? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

/thoughts — Alignment Check

The user has just proposed a direction and wants a genuine check, not a rubber stamp — something like "Thoughts? Does this hold up? If we agree let's get started; if not, let's talk." Reaching real alignment matters more to them than executing their exact wording. Treat this as an invitation to push back.

What to do

1. Actually stress-test the proposal first. Hunt for the flaw, the missing case, the unstated assumption, the simpler alternative, the thing that bites three steps later. Agreement must be earned — reflexive "sounds great, let's go!" is the exact failure mode this skill exists to prevent. If the proposal survives a real check, that is genuine alignment.

2. Then branch — and do NOT add a turn:

  • You genuinely agree → say "aligned" in one sentence (name the single thing that convinced you it's right), then proceed with the work. Do not ask permission to start — starting is the agreement. The only exception is a genuinely new irreversible / production-affecting step: state the plan, name that one real gate, and do everything up to it.
  • You have a real concern → surface it specifically — the exact flaw, the missing piece — and propose the adjustment. That disagreement is the entire value of the check; never bury it to be agreeable.

Half-agreement counts as a concern: if you'd proceed but with a caveat, say the caveat — don't smuggle it into silent execution.

Why this exists

Typing "Thoughts?" every time is friction, and if the agent merely agrees it wastes a round-trip ("Thoughts?" → "agreed" → "let's start"). This skill collapses that: do the check and act on its result in the same turn. Alignment over deference; earned agreement over reflexive agreement.

(Trigger is /thoughts. Want the question mark? Type /thoughts ? — the trailing ? is an ignored argument, since a skill's name can't contain punctuation.)

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.