Get Unstuck

A recovery guide for when a coding agent or its operator is stuck. It pauses the work, summarizes what has happened, lists unanswered questions, and suggests a next step.

In plain words
What is it for?
Use it during a stalled coding task, repeated failing fixes, unclear plans, or uncertainty about whether work is ready to commit.
Why use it?
It reduces repeated attempts and analysis paralysis when tests, edits, or decisions are going nowhere.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/cocorof/geny-executor/stuck
Any agent
npx skills add CocoRoF/geny-executor --skill stuck
Clone the repo
git clone --depth 1 https://github.com/CocoRoF/geny-executor

Made for: Claude Code, Codex.

Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 767 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00038 $0.00767
Opus 5 $0.00019 $0.00383
Sonnet 5 $0.00008 $0.00153
Haiku 4.5 $0.00004 $0.00077

Measured 2d ago against content hash a56c4599e6d2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

Get Unstuck scanned grade A with 0 findings 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 2d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

src/geny_executor/skills/bundled/stuck/SKILL.md · 91 lines

How it starts

The opening of the file, as written. The whole thing — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Stuck — recovery checklist

Stop. Don't make another tool call. Read this whole skill before deciding what to do next.

What "stuck" actually looks like

You probably got here because one of these is true:

  • The same file keeps being edited and reverted, or the same tool keeps being called with slightly different inputs and the same result.
  • You've explained a plan three different ways and the user is still unhappy.
  • A test / build keeps failing and your fixes aren't moving the failure mode, just relocating it.
  • You're about to open a PR / send a commit but you can't summarise what changed.

If none of those describe the moment, this skill probably isn't what you needed.

Recovery — five-step recipe

1. State the original goal in one sentence

What was the user's original ask? Not your interpretation, not the third-derivative subgoal you ended up chasing. The literal prompt or its plain-English rephrasing.

2. List what's actually been done

Not what you tried, not what you intended. What's committed to disk or visible in the conversation right now. Bullet list of ≤ 5 items.

3. List the open question(s)

What single piece of information would unblock you? It's almost always one of:

  • A clarification from the user (which of two interpretations, what counts as "done", which file is canonical).
  • A reading you skipped (a config you haven't opened, a log line you haven't grepped, a manifest you assumed about).
  • A test / repro you don't have (you're pattern-matching on a bug report instead of running the code).

4. Propose one concrete next step

Pick the smallest action that produces new information. Not a plan, not a refactor — one Read, one Bash command, one user question. The goal is to break the symmetry of being stuck.

5. Surface a "stop here?" prompt

End with a single yes/no question to the user. Examples:

  • "I think the issue is X. Should I try Y, or did you mean Z?"
  • "Before I keep editing, can you confirm the file at PATH is the right one?"
  • "I've made N changes; should I commit what's there now and continue from a clean state?"

Read the full file on GitHub · 91 lines

Changes

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.

  1. 2d ago First seen · 91 lines · 38 tokens per session scan A a56c4599e6d2

Subscribe to this mod's changes

Get Unstuck is a skill published in the GitHub repository CocoRoF/geny-executor (2 stars, last pushed 4d ago), licensed Apache-2.0. It adds 38 tokens to every session and 767 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.