Agent skill

Urgent Help Escalation Call

by CALLE-AI in CALLE-AI/awesome-phone-call-agents

When a vulnerable person asks for help inside an app, ring the human responsible for them and say only what is needed — who asked, and when.

MITAuto-check passed

Install Urgent Help Escalation Call

skills CLI
$ npx skills add CALLE-AI/awesome-phone-call-agents --skill urgent-help-escalation-call -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install CALLE-AI/awesome-phone-call-agents urgent-help-escalation-call --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/CALLE-AI/awesome-phone-call-agents.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/urgent-help-escalation-call .claude/skills/urgent-help-escalation-call && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
urgent-help-escalation-call
GitHub stars
107
Token cost
~2.3k tokens
SKILL.md length
1,280 words
Files
9 (incl. scripts, references)
Skills in repo
101
Repo updated
First seen
Licence
MIT

At a glance

When a vulnerable person asks for help inside an app, ring the human responsible for them and say only what is needed — who asked, and when.

  • Works in 7 steps: Confirm the trigger is genuinely urgent… → Check the cooldown for this subject. If… → Build the message from… → …
  • Panic and help buttons
  • SKILL.md covers The two people, When To Use, When Not To Use and Required Fields, plus 8 more sections
  • Runs JavaScript scripts from its folder

What it does

Urgent Help Escalation Call is an agent skill from CALLE-AI/awesome-phone-call-agents. When a vulnerable person asks for help inside an app, ring the human responsible for them and say only what is needed — who asked, and when. Returns whether the message actually landed and whether anyone confirmed they are going. Use for panic and help buttons, lone-worker duress alarms, fall and inactivity alerts, post-discharge escalation, care-home call bells, and any "the notification is not enough, phone someone" workflow.

Its SKILL.md is about 2.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 10 other files, including scripts and reference files (for example `references/call-goal.md`, `references/consent-and-boundaries.md` and `references/escalation-ladder.md`).

The repository describes itself as: Portable phone-call Agent Skills, apps, examples, adapters, and scheduler recipes for AI agents. The licence is MIT.

When your agent uses it

  • Panic and help buttons
  • Lone-worker duress alarms
  • Fall and inactivity alerts
  • Post-discharge escalation

Example prompts

  • “the notification is not enough, phone someone”
  • “/urgent-help-escalation-call”

Requirements

  • Node.js

Workflow steps

7 steps, taken from the first numbered list in SKILL.md.

  1. Confirm the trigger is genuinely urgent and came from the subject or a
  2. Check the cooldown for this subject. If a call went out recently, stop —
  3. Build the message from references/call-goal.md.
  4. Place exactly one call to recipient_phone, with an idempotency key derived
  5. Read the structured result: did it reach the recipient, and did they confirm.
  6. If it did not reach them, follow
  7. Record the outcome where the humans involved can see it.

What it can do on your machine

Read from SKILL.md and the folder at commit 38d4118. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (JavaScript), which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Urgent Help Escalation Call loads about 2.3k tokens when it runs, and up to ~7.4k if it reads all its reference files. Until then it costs about 115 tokens; SKILL.md has 1,280 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~115
When it runs · the whole SKILL.md, loaded when a task matches
~2.3k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~7.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from CALLE-AI/awesome-phone-call-agents at commit 38d4118, republished under its MIT licence (© CALLE-AI). 1,280 words, ~2,346 tokens.

Download SKILL.mdSave it as .claude/skills/urgent-help-escalation-call/SKILL.md (or your agent's skills folder). This skill also uses 8 other files; get the full folder from GitHub.
name
urgent-help-escalation-call
description
When a vulnerable person asks for help inside an app, ring the human responsible for them and say only what is needed — who asked, and when. Returns whether the message actually landed and whether anyone confirmed they are going. Use for panic and help buttons, lone-worker duress alarms, fall and inactivity alerts, post-discharge escalation, care-home call bells, and any "the notification is not enough, phone someone" workflow.
license
MIT

Urgent Help Escalation Call

Use this skill when software has learned that someone needs a human, and a notification is not good enough.

An in-app alert is only as good as somebody looking at a screen. The moments when that assumption fails are exactly the moments the alert exists for: the middle of the night, a shower, a meeting, a commute. A ringing phone is the one channel that reliably interrupts.

This skill is the bridge. One trigger, one call, one clear answer about whether it landed.

The two people

Getting this wrong is the failure mode that matters most, so name them separately in your own code:

WhoRole
Subjectthe vulnerable personasked for help; is never called by this skill
Recipientthe carer, responder, or next of kinthe one whose phone rings

Dialling the subject instead of the recipient means telling a frightened person that a frightened person needs help. Keep the two numbers in separate fields, label them unambiguously in any UI, and validate them separately.

When To Use

  • panic buttons and help buttons in apps used by vulnerable people
  • lone-worker duress alarms
  • fall detection, inactivity timeouts, or bed-exit alerts
  • care-home call bells that need to reach staff who are not at a station
  • post-discharge deterioration flags
  • any in-app urgent signal where "they will see the notification" is a bad bet

When Not To Use

Do not use this skill to:

  • replace emergency services — see Emergencies
  • call the subject themselves, for any reason
  • deliver clinical detail, triage, or advice by phone
  • send routine or digest notifications; this is for urgent signals only
  • escalate something the subject did not actually ask for, without saying so
  • call a responder who has not agreed to be called

Required Fields

Ask for anything missing. Never infer these:

FieldNotes
recipient_phoneE.164. The responder's number, never the subject's. Validate; never guess a country code.
subject_nameWhat the responder should be told, e.g. "Eleanor".
triggered_atWhen the request happened, so the call can say when rather than "just now".
recipient_nameOptional but strongly preferred — being addressed by name at 3am is worth a lot.
regionOptional. Chosen, not inferred — it decides which line the call comes from.

Core Workflow

  1. Confirm the trigger is genuinely urgent and came from the subject or a sensor, not from a routine event.
  2. Check the cooldown for this subject. If a call went out recently, stop — see references/repeat-triggers.md.
  3. Build the message from references/call-goal.md. Short. One fact, one instruction.
  4. Place exactly one call to recipient_phone, with an idempotency key derived from the recipient and the trigger minute so a retried request cannot ring twice.
  5. Read the structured result: did it reach the recipient, and did they confirm.
  6. If it did not reach them, follow references/escalation-ladder.md.
  7. Record the outcome where the humans involved can see it.

Fire the call without blocking the trigger. The subject is standing there having asked for help; the app's response to them must not wait on a phone network, and a failure to connect must never surface to them as an error. Raise your in-app alert first, place the call alongside it.

What The Call Says

Confirm who is listening before you disclose anything. This ordering is the design, not a nicety:

"Hello, this is <product>. Am I speaking with <recipient>?"
   ├─ no / unsure → "Sorry to have troubled you. Goodbye."  → hang up, say nothing
   └─ yes         → "<subject> has pressed the help button and is asking for
                     you. Please go to them when you can."
                  → "Are you able to get to them?" → thank them → hang up

The obvious design leads with the message to save a second. It is wrong. A number will eventually be answered by a neighbour, a partner, a child, or a stranger with a recycled number — and leading with the message tells them a named vulnerable person has asked for help before you know who is listening. Nothing downstream can un-say it, and a "if it's the wrong person, stop" branch fires only after the model has already spoken.

After the gate:

  • Do not sound alarmed. Panic on the phone gets nobody there faster.
  • Do not speculate about why. You do not know, and a guess becomes what the responder believes on the drive over.
  • Say nothing about the subject's health, condition, or history.
  • Never ask a stranger to pass the message on — that is a disclosure with extra steps.
  • Ask whether they can go — a question, not an instruction. "Confirm you are going to her" reads as an order barked at someone just woken up. Ask once, accept the answer, do not press. They may be two hours away.
  • Sound like a person passing on a message, not a system issuing one. Models drift officious when handed numbered steps and a schema; say the register you want out loud.
  • Voicemail names nobody. It is the one case where the gate cannot run, so the message must be safe for a whole room to overhear: no name, no reason, just "there is an alert waiting in your app".
  • Keep it under a minute.

Full template and the reasoning line by line: references/call-goal.md.

Show full SKILL.md (474 more words)Show less

Structured Result

Two questions, no more:

json
{
  "reached": "caregiver | someone_else | voicemail | no_answer",
  "acknowledged": "yes | no | unknown"
}

reached is not the same as acknowledged. A voicemail is a message left, not a person on their way, and the difference decides whether the ladder continues. Full schema and the reasoning: references/result-schema.md.

Simpler than a call to a vulnerable person, but not nothing:

  • The recipient agreed to be called. They configured the number, or someone authorized to did. Record who and when.
  • The subject knows this happens. Pressing a help button should be understood as "this will fetch someone", and any consent flow should say so plainly rather than burying it.
  • Health information is disclosed to whoever answers. That is why the message says only that someone asked for you. See references/consent-and-boundaries.md.

Repeat Triggers

Someone frightened, confused, or in pain may press a button many times in a minute. The responder needs one call, not eleven — and each one costs money.

Hold a cooldown per subject, not per request. Suppress inside the window and log that you did. Never let one subject's cooldown silence another's. Details and the trade-offs: references/repeat-triggers.md.

Emergencies

This is not an emergency service and must never be presented as one.

It reaches one nominated person, who may be asleep, driving, or unreachable. It cannot assess anything. It does not know whether the subject is hurt.

Say plainly in your product and your setup docs that the subject must have a real route to emergency help that does not depend on this call being placed, answered, or correctly understood.

Side Effects, Cancellation, Cost

  • Side effect: one real outbound phone call to a real person, per trigger that survives the cooldown.
  • Cancellation: nothing to cancel once a call connects. This skill creates no provider-side recurrence — it fires on an event, not a schedule.
  • Cost: one call from the CALL-E account per escalation. The cooldown is what stands between a distressed subject and a large bill.
  • No-call path: this contribution ships a fake CALL-E server (scripts/fake-calle-api.mjs) serving the documented shapes, so the trigger, call contract, structured result, and escalation ladder can be exercised without ringing anyone. Use it for tests, CI, and demos.

Reference design

This pattern was developed for Yadira, a companion for people living with dementia. In that product, a person presses a help button; the ordinary in-app alert is raised immediately, and CALL-E separately rings the configured caregiver. The subject's number is never supplied to this workflow.

Two decisions from that implementation are worth copying:

Do not make the phone call the only alert. A call can be delayed, silenced, or missed. Raise the in-app alert first and treat the phone call as an additional interruptive channel.

The provider's voice is not yours. If your product depends on a specific, familiar voice for the vulnerable person, an outbound provider voice may be the wrong channel. This skill deliberately calls the responder instead.

© CALLE-AI, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 8 other files (scripts, references) in skills/urgent-help-escalation-call of CALLE-AI/awesome-phone-call-agents.

  • SKILL.md
  • references/call-goal.md
  • references/consent-and-boundaries.md
  • references/escalation-ladder.md
  • references/examples.md
  • references/repeat-triggers.md
  • references/result-schema.md
  • references/safety.md
  • scripts/fake-calle-api.mjs

Open the folder on GitHubat commit 38d4118

Compare with similar skills

Urgent Help Escalation Call next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Urgent Help Escalation Call compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Urgent Help Escalation Call this skillCALLE-AI/awesome-phone-call-agents107—~2.3kAutomated safety check: PassMIT
Vulnerability Scanningsickn33/agentic-awesome-skills47k1 repos~2.8kAutomated safety check: PassMIT
Performing Vulnerability Scanning With Nessusmukul975/Anthropic-Cybersecurity-Skills34k—~2.8kAutomated safety check: PassApache-2.0
Building Vulnerability Scanning Workflowmukul975/Anthropic-Cybersecurity-Skills34k—~3.2kAutomated safety check: PassApache-2.0
Performing Ot Vulnerability Assessment With Clarotymukul975/Anthropic-Cybersecurity-Skills34k—~3.1kAutomated safety check: PassApache-2.0
Cra Vulnerability Obligationsdavila7/claude-code-templates33k1 repos~4.5kAutomated safety check: PassCC-BY-4.0

Similar skills

  • Vulnerability Scanning

    sickn33/agentic-awesome-skills

    Scan systems and dependencies for CVEs and security vulnerabilities.

    47k GitHub starsUsed in 1 repo~2.8k tokens
    SecurityAuto-check passed
  • Performing Vulnerability Scanning With Nessus

    mukul975/Anthropic-Cybersecurity-Skills

    Performs authenticated and unauthenticated vulnerability scanning using Tenable Nessus to identify known vulnerabilities, misconfigurations, default credentials, and missing patches across network…

    34k GitHub stars~2.8k tokensUpdated 1 mo ago
    SecurityAuto-check passed
  • Building Vulnerability Scanning Workflow

    mukul975/Anthropic-Cybersecurity-Skills

    Builds a structured vulnerability scanning workflow using tools like Nessus, Qualys, and OpenVAS to discover, prioritize, and track remediation of security vulnerabilities across infrastructure.

    34k GitHub stars~3.2k tokensUpdated 1 mo ago
    SecurityAuto-check passed
  • Performing Ot Vulnerability Assessment With Claroty

    mukul975/Anthropic-Cybersecurity-Skills

    Perform OT vulnerability assessments using the Claroty xDome platform for asset discovery, risk scoring, and vulnerability correlation, combining passive traffic-based identification and active safe…

    34k GitHub stars~3.1k tokensUpdated 1 mo ago
    SecurityAuto-check passed
  • Cra Vulnerability Obligations

    davila7/claude-code-templates

    A skill your agent uses when a user asks what the EU Cyber Resilience Act (CRA) means for their product, whether and when they must report a vulnerability or incident, or what a specific CVE…

    33k GitHub starsUsed in 1 repo~4.5k tokens
    SecurityAuto-check passed
  • Gathers security findings from Dependabot, GCP container scanning, Docker Scout and Linear security issues, then triages and remediates them across Warp's repos and images.

    65k GitHub starsUsed in 1 repo~2.1k tokens
    SecurityAuto-check passed

More from CALLE-AI/awesome-phone-call-agents

All 101 skills in this repo
  • Accessible Outing Verifier

    CALLE-AI/awesome-phone-call-agents

    Demonstrates advisory accessibility-planning checks with offline fixtures and a proposed bounded CALL-E workflow; use for exploring unknown or qualified venue claims without making calls.

    107 GitHub stars~1.4k tokensUpdated yesterday
    Auto-check passed
  • Ground Truth Gate

    CALLE-AI/awesome-phone-call-agents

    A skill your agent uses when an agent holds some evidence for a physical-world claim but the evidence is broader, narrower, or older than the exact question asked, and it must first decide whether a…

    107 GitHub stars~3.3k tokensUpdated yesterday
    Auto-check passed
  • Is It Accessible

    CALLE-AI/awesome-phone-call-agents

    Call a venue and ask the accessibility questions that matter to one specific person — step-free entry, hearing loop, guide dogs, quiet hours, changing places — then return a per-need verdict backed…

    107 GitHub stars~4.3k tokensUpdated yesterday
    Auto-check passed
  • Landmark Navigation Assist

    CALLE-AI/awesome-phone-call-agents

    Turns a pre-written, building-level location config into a CALL-E outbound phone-call task that guides a delivery driver through the last few hundred metres to a specific building using landmarks…

    107 GitHub stars~1.6k tokensUpdated yesterday
    Auto-check passed
  • Research Gap Call Verifier

    CALLE-AI/awesome-phone-call-agents

    Turn cited business research into a bounded, approval-gated phone-call plan that asks only unresolved factual questions, then reconcile CALL-E-compatible results without treating voicemail, refusal…

    107 GitHub stars~1.7k tokensUpdated yesterday
    Auto-check passed
  • Structured Outcome Followup Call

    CALLE-AI/awesome-phone-call-agents

    Place a goal-driven CALL-E call that collects specific structured answers, score those answers against a deterministic rubric you supply, and conditionally trigger a follow-up action — all runnable…

    107 GitHub stars~1.4k tokensUpdated yesterday
    Auto-check passed

Questions about Urgent Help Escalation Call

What does Urgent Help Escalation Call do?

When a vulnerable person asks for help inside an app, ring the human responsible for them and say only what is needed — who asked, and when. Urgent Help Escalation Call is an agent skill from CALLE-AI/awesome-phone-call-agents. When a vulnerable person asks for help inside an app, ring the human responsible for them and say only what is needed — who asked, and when.

When should I use Urgent Help Escalation Call?

Urgent Help Escalation Call fits situations like: panic and help buttons; lone-worker duress alarms; fall and inactivity alerts; post-discharge escalation.

How do I install Urgent Help Escalation Call in Claude Code?

Run `npx skills add CALLE-AI/awesome-phone-call-agents --skill urgent-help-escalation-call -a claude-code`. Or copy the skill folder (skills/urgent-help-escalation-call in CALLE-AI/awesome-phone-call-agents) into .claude/skills/urgent-help-escalation-call in your project. Claude Code loads it when a task matches its description.

How do I install Urgent Help Escalation Call in Codex?

Run `npx skills add CALLE-AI/awesome-phone-call-agents --skill urgent-help-escalation-call -a codex`. Or copy the skill folder (skills/urgent-help-escalation-call in CALLE-AI/awesome-phone-call-agents) into .agents/skills/urgent-help-escalation-call in your project. Codex loads it when a task matches its description.

Can I use Urgent Help Escalation Call in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add CALLE-AI/awesome-phone-call-agents --skill urgent-help-escalation-call -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/urgent-help-escalation-call, .gemini/skills/urgent-help-escalation-call, .github/skills/urgent-help-escalation-call and .opencode/skills/urgent-help-escalation-call in your project.

What does Urgent Help Escalation Call need to run?

Going by SKILL.md and its folder, Urgent Help Escalation Call needs JavaScript for the scripts in its folder. Our summary lists: Node.js.

Does Urgent Help Escalation Call access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Urgent Help Escalation Call safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Urgent Help Escalation Call use?

Urgent Help Escalation Call is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Urgent Help Escalation Call use?

About 2.3k tokens (SKILL.md is roughly 9.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 5k tokens, read only when the agent opens those files.

What are the alternatives to Urgent Help Escalation Call?

Skills that share tags, products or a category with Urgent Help Escalation Call: Vulnerability Scanning (sickn33/agentic-awesome-skills, 47k stars), Performing Vulnerability Scanning With Nessus (mukul975/Anthropic-Cybersecurity-Skills, 34k stars), Building Vulnerability Scanning Workflow (mukul975/Anthropic-Cybersecurity-Skills, 34k stars) and Performing Ot Vulnerability Assessment With Claroty (mukul975/Anthropic-Cybersecurity-Skills, 34k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Urgent Help Escalation Call?

CALLE-AI (a GitHub organization) maintains it in CALLE-AI/awesome-phone-call-agents, which has 107 GitHub stars. The repository holds 101 skills in this directory. The repository was last updated on October 10, 2026.

Source: CALLE-AI/awesome-phone-call-agents on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.