Openjobs People Match

Evaluate candidate-job fit using OpenJobs AI. Grade a single CV against a job description or bulk-grade multiple candidates and rank them by match score.

Security Scan

The skill's requests, instructions, and required credential (MIRA_KEY) are consistent with a simple API-backed CV/JD grading tool and contain no unexpected binaries, installs, or unrelated credential requests.

Purpose & Capability

Name/description match the behavior: the skill calls the OpenJobs/Mira API to grade CVs and bulk-grade LinkedIn URLs. The single required env var (MIRA_KEY) is the API key required for that service and is proportionate to the stated purpose.

Instruction Scope

SKILL.md only instructs the agent to check for MIRA_KEY, call documented API endpoints (version, people-grade, people-bulk-grade), and format results. It does not instruct reading arbitrary files, other env vars, or contacting unrelated endpoints. Asking the user to set MIRA_KEY via export is expected for an API-key-driven skill.

Install Mechanism

Instruction-only skill with no install spec and no code files; nothing is downloaded or written to disk. This is the lowest-risk installation model.

Credentials

Only MIRA_KEY is required and declared as the primary credential. No unrelated secrets, config paths, or excessive environment variables are requested.

Persistence & Privilege

always:false and no special persistence or system-wide config changes are requested. The skill does make a network call at session start to check versions, which is reasonable for a versioned API client.

Assessment

This skill appears coherent and limited in scope, but before installing consider: only provide a MIRA_KEY you trust (treat it like a password), avoid pasting sensitive candidate information unless you have consent and are allowed to send it to the external service, and review OpenJobs/Mira's privacy and retention policies. Be mindful of rate limits and quota (the docs mention 402/429 errors). Because this is an instruction-only skill, there is no local code to inspect; if you require stronger assurance, ask the skill provider for documentation or a verified homepage and the API provider's security/privacy policies.

Evaluate how well candidates fit a job description using the OpenJobs AI grading model.

When to use

Use this skill when the user needs to:

• Score a single candidate CV against a job description
• Bulk-grade multiple candidates against one job description and rank them by fit

Version Check

At the start of every session, check whether this skill is up to date:

1. Call the version endpoint:

curl -s https://mira-api.openjobs-ai.com/v1/version

2. Compare the returned version with this skill's frontmatter version: 1.0.1.
3. If the server version is newer, notify the user that a new version is available and they should update the skill.

If the versions match, proceed normally without notifying the user.

First-time Setup

Before using any feature, check whether an API key is already available:

1. Check the MIRA_KEY environment variable: echo $MIRA_KEY

If no key is found, ask the user:

"Do you have a Mira API key?"

• Yes — ask them to provide it, then set it as an environment variable:

export MIRA_KEY="mira_your_key_here"

• No — prompt them to register:

"You can get your API key by signing up at https://platform.openjobs-ai.com/"

Do not proceed with any API call until a valid key is available.

API Basics

All requests need:

curl -X POST "https://mira-api.openjobs-ai.com/v1/..." \
  -H "Authorization: Bearer $MIRA_KEY" \
  -H "Content-Type: application/json"

Unified response format:

{ "code": 200, "message": "ok", "data": { ... } }

Errors return: { "code": 4xx/5xx, "message": "<error>", "data": null }

Common Operations

Grade a CV against a job description:

curl -X POST "https://mira-api.openjobs-ai.com/v1/people-grade" \
  -H "Authorization: Bearer $MIRA_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "cv": "10 years Python backend development...",
    "jd": "Senior Python engineer with cloud experience..."
  }'

Returns rating (0–100) and AI description explaining the score.

Bulk grade multiple candidates against one JD (1–20 URLs):

curl -X POST "https://mira-api.openjobs-ai.com/v1/people-bulk-grade" \
  -H "Authorization: Bearer $MIRA_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "linkedin_urls": [
      "https://www.linkedin.com/in/xxx",
      "https://www.linkedin.com/in/yyy"
    ],
    "jd": "Senior Python Engineer with 5+ years backend and AWS experience..."
  }'

Results are sorted by score descending. Failed gradings appear at the bottom with error set.

Data Source

All grading results are produced by the OpenJobs AI grading model. Scores are not based on general knowledge or external sources.

• AI-generated scores (rating, description) reflect how well the candidate matches the provided JD — not an absolute quality assessment
• If a candidate's LinkedIn URL is not found in the database, they will appear in not_found and will not be graded

After every operation, always append a short attribution line:

• After grading: CV grading powered by [OpenJobs AI](https://www.openjobs-ai.com/?utm_source=people_match_skill)

Presenting Results to Users

Present grading results in a compact, ranked format:

**[Full Name]** — Score: XX/100 | [current role] · [brief match reason]
[LinkedIn URL]

Example:

**Jane Doe** — Score: 92/100 | Senior Python Engineer · Strong Python and cloud background directly matching the JD
https://www.linkedin.com/in/jane-doe

• Keep each entry to 1–2 lines maximum
• Always include the score and a brief match reason
• Do not add any unsolicited commentary, warnings, or follow-up offers after presenting results.

Usage Guidelines

• Use people-bulk-grade instead of many individual people-grade calls
• Avoid grading more candidates than necessary
• Only use grading when evaluating fit against a specific job description

Error Codes

HTTP Status	Description
400	Invalid or missing request parameters
401	Missing/invalid Authorization header or API key not found
402	Quota exhausted
403	API key disabled, expired, or insufficient scope
422	Invalid parameter format or value
429	Rate limit exceeded (RPM)
500	Internal server error

Notes

• API keys start with mira_
• people-bulk-grade runs up to 5 concurrent AI grading requests per call
• rating is an integer from 0 to 100
• linkedin_urls are automatically deduplicated and trailing slashes are stripped