# hh-openclaw-agent

Adopt this skill: https://proskills.md/skills/other/zack-dev-cm-hh-openclaw-agent-skill-hh-openclaw-agent

Page: https://proskills.md/skills/other/zack-dev-cm-hh-openclaw-agent-skill-hh-openclaw-agent
This document is the public skill bundle for that page. It contains SKILL.md and every helper file the catalog carries. No login is required.

## File: skill/hh-openclaw-agent/SKILL.md
size: 4515
upstream: https://github.com/zack-dev-cm/hh-openclaw-agent/blob/main/skill/hh-openclaw-agent/SKILL.md

```text
---
name: hh-openclaw-agent
description: HH OpenClaw Agent is a public ClawHub hh.ru application skill. Use it when the user says "hh openclaw agent", "hh job apply automation", "hh.ru browser application", or wants a reviewed hh.ru application packet with logged-in browser execution and an auditable submission bundle.
version: 1.0.7
homepage: https://github.com/zack-dev-cm/hh-openclaw-agent
license: MIT-0
user-invocable: true
metadata: {"openclaw":{"homepage":"https://github.com/zack-dev-cm/hh-openclaw-agent","skillKey":"hh-openclaw-agent","requires":{"anyBins":["python3","python"]}}}
---

# HH OpenClaw Agent

Search intent: `hh openclaw agent`, `hh.ru apply`, `job application browser automation`, `career submission bundle`

## Goal

Turn one `hh.ru` application run into a reusable submission bundle:

- one machine-readable application packet
- one ordered execution log with evidence
- one structural bundle check
- one shareable markdown report

This skill is for reviewed `hh.ru` browser execution through OpenClaw.
It assumes the browser profile is already authenticated.

## Use This Skill When

- the user wants to prepare and submit an `hh.ru` application through OpenClaw
- a vacancy-specific packet should be reviewed before the live browser write
- the same `hh.ru` response flow needs an auditable run log instead of chat memory
- a blocked or failed submission needs screenshots, exact steps, and a handoff report
- the operator wants one record of approved content and live browser outcome

## Quick Start

1. Initialize the application packet.
   - Use `python3 {baseDir}/scripts/init_hh_application_packet.py --out <json> --packet-id <id> --vacancy-id <id> --vacancy-title <title> --company-name <company> --resume-title <resume> --vacancy-url <url> --apply-url <url> --cover-letter <text>`.
   - Add `--browser-profile`, `--review-status`, `--reviewer`, and repeatable `--blocked-action` fields when needed.

2. Execute the browser flow through OpenClaw.
   - Open the apply URL in a logged-in profile.
   - Record each meaningful browser action with `append_hh_execution_step.py`.
   - Capture screenshots and outcome notes as the run progresses.

3. Keep operator-owned auth and approval gates explicit.
   - If `hh.ru` shows login, CAPTCHA, 2FA, or another auth challenge, stop and let the operator complete it in the same browser profile.
   - Do not live-send a packet that is still `pending` or `rejected`.

4. Check the bundle before sharing or counting it as complete.
   - Use `python3 {baseDir}/scripts/check_hh_submission_bundle.py --manifest <json> --repo-root <repo> --out <json>`.
   - Fix approval-gate issues, missing screenshots, or incomplete failed-step notes before final handoff.

5. Render the report.
   - Use `python3 {baseDir}/scripts/render_hh_submission_report.py --manifest <json> --out <md>`.
   - The default report redacts the cover letter body, keeps only public `https://*.hh.ru/...` URLs, and redacts private artifact paths.
   - Add `--include-sensitive` only when you intentionally need a full-content export.
   - Share the report instead of loose screenshots and manual notes.

## Operating Rules

### Packet rules

- Keep one packet per vacancy response flow.
- The cover letter should stay vacancy-specific.
- Record vacancy URL, apply URL, resume title, and approval state before opening the live form.

### Execution rules

- Use a logged-in OpenClaw browser profile for the live `hh.ru` session.
- Record expected result and actual result for every meaningful step.
- Capture a screenshot for failed or blocked steps and for the final submitted state when possible.
- Keep artifact paths relative so the bundle can move between machines.

### Safety rules

- Do not claim undocumented `hh.ru` write APIs.
- Do not store cookies, secrets, or tokens in notes or artifacts.
- Do not send a live application unless the packet review state is `approved`.
- Treat login, CAPTCHA, passkey, and 2FA as operator-owned interruptions, not background automation.

## Bundled Scripts

- `scripts/init_hh_application_packet.py`
  - Create a machine-readable application packet for one `hh.ru` response flow.
- `scripts/append_hh_execution_step.py`
  - Append one evidence-backed browser execution step to the packet.
- `scripts/check_hh_submission_bundle.py`
  - Validate approval state, execution evidence, and bundle safety before handoff.
- `scripts/render_hh_submission_report.py`
  - Render a concise markdown report from the packet and execution log.
```

## File: README.md
upstream: https://github.com/zack-dev-cm/hh-openclaw-agent/blob/main/README.md

````text
# HH OpenClaw Agent

**Prepare a reviewed `hh.ru` application packet, execute the browser flow through OpenClaw, and capture one auditable submission bundle.**

HH OpenClaw Agent is a small public OpenClaw skill for `hh.ru` application work. It creates
an application packet, records OpenClaw execution steps, checks the resulting submission bundle, and renders
a shareable markdown report for review, debugging, or funnel tracking.

## Proof

```md
# HH Submission Report

- Apply URL: https://hh.ru/applicant/vacancy_response
- Review status: **pending**

## Cover Letter

_Redacted by default._

### 1. open-response-flow
- Artifacts: screenshot: `artifacts/response-form.png`
```

## Quick Start

```bash
python3 skill/hh-openclaw-agent/scripts/init_hh_application_packet.py \
  --out /tmp/hh-packet.json \
  --packet-id hh-demo-001 \
  --vacancy-id 135791357 \
  --vacancy-title "Senior ML Engineer" \
  --company-name "Example AI" \
  --resume-title "Senior ML Engineer / AI Systems" \
  --vacancy-url 'https://hh.ru/vacancy/135791357' \
  --apply-url 'https://hh.ru/applicant/vacancy_response?vacancyId=135791357' \
  --browser-profile hh-main \
  --cover-letter "Здравствуйте. Мне интересна эта роль, потому что у меня сильный практический опыт в production ML и AI delivery."

python3 skill/hh-openclaw-agent/scripts/append_hh_execution_step.py \
  --manifest /tmp/hh-packet.json \
  --step-id open-response-flow \
  --action "Open the direct vacancy response page in a logged-in profile" \
  --expected "The response form loads and the cover letter field is visible" \
  --actual "The response form opened normally" \
  --status passed \
  --effect inspected \
  --screenshot artifacts/response-form.png

python3 skill/hh-openclaw-agent/scripts/check_hh_submission_bundle.py \
  --manifest /tmp/hh-packet.json \
  --repo-root . \
  --out /tmp/hh-packet-check.json

python3 skill/hh-openclaw-agent/scripts/render_hh_submission_report.py \
  --manifest /tmp/hh-packet.json \
  --out /tmp/hh-packet-report.md
```

Quote URLs with query strings when running the example in `zsh`.
The markdown report is redacted by default. Use `--include-sensitive` only when you intentionally need a full export.
Default shareable output keeps only public `https://*.hh.ru/...` URLs and redacts private artifact paths.

## What It Covers

- one machine-readable application packet for a single `hh.ru` response flow
- evidence-backed execution steps for review, browser interaction, submit, and outcome capture
- bundle validation for approval-gate mistakes, missing failure detail, and absolute artifact paths
- a shareable markdown report for review, debugging, or job-funnel records, with sensitive fields redacted by default

## Included

- `skill/hh-openclaw-agent/SKILL.md`
- `skill/hh-openclaw-agent/agents/openai.yaml`
- `skill/hh-openclaw-agent/scripts/init_hh_application_packet.py`
- `skill/hh-openclaw-agent/scripts/append_hh_execution_step.py`
- `skill/hh-openclaw-agent/scripts/check_hh_submission_bundle.py`
- `skill/hh-openclaw-agent/scripts/render_hh_submission_report.py`

## Use Cases

- prepare a vacancy-specific packet before opening the live `hh.ru` browser flow
- keep one auditable record of what was approved, what was submitted, and what blocked the run
- debug an `hh.ru` response flow with screenshots and exact browser steps
- hand a live application run to another operator without losing the context

## License

MIT No Attribution (MIT-0)
````
