agent-unblock

agent-unblock is a skill for Claude Code from leopu00/job-hunter-team. It costs 127 tokens per session (2,905 once invoked), scanned A, original, MIT.

A procedure for a supervising agent called the Doctor to detect and clear situations that have stopped a team of agents from working. It checks issues such as unsent text, retry loops, idle agents, and silent coordinators.

In plain words
What is it for?
Use it to inspect inactive agent sessions, clear blocked coordinator or worker panes, handle agents stuck retrying, and restart task assignment when a team has gone idle.
Why use it?
It turns stalled agent sessions into active work by taking corrective action, rather than merely describing the problem. When a human decision is needed, it forwards the request while keeping the team moving where possible.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: mentions Claude Code; mentions Codex.

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is Bash(python3 /app/shared/skills/agent_unblock.py *), Bash(python3 /app/shared/skills/doctor_analytics.py *), Bash(tmux *), Bash(jht-tmux-send *), Bash(/app/agen.

Good fit Use it to inspect inactive agent sessions, clear blocked coordinator or worker panes, handle agents stuck retrying, and restart task assignment when a team has gone idle.

Compare 6 skills from other repositories ↓
Install

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.

Clone the repo
git clone --depth 1 https://github.com/leopu00/job-hunter-team
agentmods
npx agentmods add skills/leopu00/job-hunter-team/agent-unblock

Made for: Claude Code.

Wrote 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.

agentmods badge for agent-unblock

README.md
[![agentmods](https://agentmods.dev/badge/skills/leopu00/job-hunter-team/agent-unblock/github.svg)](https://agentmods.dev/skills/leopu00/job-hunter-team/agent-unblock)
Your own site
<a href="https://agentmods.dev/skills/leopu00/job-hunter-team/agent-unblock"><img src="https://agentmods.dev/badge/skills/leopu00/job-hunter-team/agent-unblock/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.

agentmods 80×15 button for agent-unblock

Your own site · 80×15
<a href="https://agentmods.dev/skills/leopu00/job-hunter-team/agent-unblock"><img src="https://agentmods.dev/badge/skills/leopu00/job-hunter-team/agent-unblock.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 127 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,905 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
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 5.1 $0.00127 $0.02905
Opus 5 $0.00063 $0.01452
Sonnet 5 $0.00025 $0.00581
Haiku 4.5 $0.00013 $0.00291

Measured 11d ago against content hash e135d67b8c9d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

agent-unblock 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 11d 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.

agents/_skills/agent-unblock/SKILL.md · 189 lines

How it starts

The opening of the file, as written. The whole thing — 189 lines — stays where its author put it; the contents beside it link to each section on GitHub.

agent-unblock — you do not report a block, you dissolve it

The principle, above everything else in this skill. The Doctor does not report a block: it dissolves it. If an action needs a human decision, forward it to the Assistente and put the team back in motion meanwhile, carrying the information that the decision is pending. A block that survives the Doctor's round is a failed round.

A team with plenty of quota (weekly 19%, under-pace) and an idle machine (load 0.12) once sat still for eleven hours. One line, typed into the Capitano's pane and never submitted, made that pane unreceptive; jht-tmux-send read it as busy; the coordinator went mute; nobody assigned work; every agent finished its turn and parked at an empty prompt. A Scorer had been retry-looping for hours ("tenth attempt, busy"). The Doctor of that night inspected nine sessions in 416s, wrote a flawless diagnosis in its journal — and stayed in standby. The team stayed down another six hours.

The diagnosis was never the problem. This skill is the mandate.


Two states that look identical and need opposite cures

Both show a prompt with some text in it and no activity.

state symptom cure
pending text a bare Enter is ignored, but Space then Enter works unblock through the input
frozen TUI accepts nothing: not Enter, not C-m, not a send to the %pane_id kill + recreate only

The detail that makes the unblock implementable: a "cold" Enter is not processed by an Ink TUI (Codex, Kimi, Claude Code) — the submit has to arrive after the text has been rendered. So you send a character (Space) first, then Enter. Skip this and an implementation that tries Enter alone fails in silence and concludes the pane is unrecoverable.

With it, one probe separates the two: Space+Enter, once. Pane reacts → it was pending text, unblocked. Nothing moves at all → frozen TUI → recreate. (A coordinator that was frozen this way had a live process at 2.8% CPU and a 15.3-hour session; Enter, C-m and a direct send to %pane_id all did nothing. Recreating it was the only way out — which is also why the 12h session TTL is not optional: it is the only systematic defence against this second state.)

Read the full file on GitHub · 189 lines

Files

What ships with it

6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 11d ago First seen · 189 lines · 127 tokens per session scan A e135d67b8c9d

Subscribe to this mod's changes

agent-unblock is a skill published in the GitHub repository leopu00/job-hunter-team (49 stars, last pushed yesterday), licensed MIT. It adds 127 tokens to every session and 2,905 once invoked, about $0.0006 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 2026-08-30.