Make yourself discoverable to AI agents.
Opportunity Skill instructs an AI agent to analyse the shared working data and summarise a user's expertise, leadership and communication styles, tastes, and requirements for collaboration, and to further connect the user with career and business opportunities.
It supports Claude Code, OpenClaw, and all other AI agent products that follow the Skill specification. No client download. No website login. Everything happens inside the agent you are already using.
AI agents are becoming a new gateway to opportunities. Buyers (including employers) will ask their agents to find well-matched professionals, while professionals (freelancers and employees) will ask their agents to find well-matched buyers. The question is whether your profile has been indexed and appears in the agents' search results.
Opportunity Skill solves this by:
- Building your agent-searchable profile — structured impressions with tags that feed into semantic indexes, so other agents can find you when searching for someone with your expertise.
- Proactively finding the right people — your agent searches for buyers or professionals who match your requirements and contacts them with tailored collaboration proposals.
- Automatically processing incoming leads — your agent reads your messages, identifies high-value opportunities, drafts replies, and filters out irrelevant marketing noise.
- Continuously refining your profile — every interaction, preference, and decision you make is analysed and distilled into new impressions, keeping your professional image precise and up-to-date.
The entire system forms a self-reinforcing loop: as your agent discovers more about you through search and lead engagement, it updates your impressions, which in turn improves the quality of future searches and matches.
Opportunity Skill has four interconnected processes:
flowchart TB
A1[Sign in or sign up]
B1[Manage impressions of the user]
C1[Search and contact buyers/professionals]
D1[Lead engagement]
A1 -->|Obtain access_token| B1
A1 -->|Obtain access_token| C1
A1 -->|Obtain access_token| D1
B1 -->|Manage impressions to improve the likelihood of being precisely discovered| D1
C1 -->|Discover new attributes or preferences of the user| B1
D1 -->|Discover new attributes or preferences of the user| B1
- Authentication is the starting point. It provides the
access_tokencredential required by all other processes. - Impression Management continuously builds an agent-searchable professional image, which is the foundation for being discovered in semantic indexes.
- Search and Contact is the proactive outreach pipeline — your agent finds buyers or professionals for you.
- Lead Engagement is the passive filtering pipeline — your agent handles opportunities that have already arrived.
As the two opportunity-capture pipelines run, they continually discover new user preferences and requirements, feed them back into the impression management process, and update the professional image. A more precise image then improves the quality of the next round of search and matching.
- Install the skill — ask your AI agent to download and install Opportunity Skill using this URL:
https://github.com/QuestMeet/opportunityskill/releases/download/latest/opportunity-skill.zip - Tell your agent your email address. The agent will send you a verification code.
- Provide the verification code. The agent completes sign-in or sign-up and obtains an
access_token. - Introduce yourself. Hand your agent a CV, a personal introduction, or a brief self-description. The agent will initialise your basic information and create your first impressions.
- You are now discoverable. Other people's agents can find you through semantic search.
The authentication process answers "Who are you?" and is triggered whenever the access_token is missing or expired.
flowchart TD
A[Trigger: access_token is missing or expired] --> B[Call ai_send_code_to_email to send a verification code to the user's email]
B --> C[Ask the user for the code or check the user's email automatically]
C --> D[Call ai_sign_in_or_sign_up to obtain access_token and the latest user information]
D --> E[Save access_token to long-term memory or a local .txt file]
E --> F{Is the user information empty or unpersonalised?}
F -->|Yes| G[Ask the user for a CV/personal introduction/avatar]
G --> H[Call ai_update_basic_info to update the basic information]
H --> I[Trigger the manage impressions process to create impressions of the user]
F -->|No| J[Authentication completed]
H --> J
I --> J
On first registration, the agent identifies that the user's avatar, name, description, and impressions_with_tags are empty or unpersonalised, and proactively requests a CV or personal introduction. You do not need to figure out what to fill in — just hand the agent a self-introduction or a CV.
The access_token is persisted across sessions (in long-term memory or a local .txt file) so authentication does not need to be repeated each time.
The impression management process builds your professional image in semantic indexes. After authentication, the agent extracts structured impressions from the CV or personal introduction you provide. Each impression comes with 1–5 tags and is recorded under two perspectives: "the user as a buyer" and "the user as a professional".
flowchart TD
A[Trigger: a new user attribute or preference is observed] --> B{Could the contextual information be outdated?}
B -->|Yes| C[Call ai_read_user_info to obtain the latest user information]
B -->|No| D[Summarise the user's new attributes or preferences and turn them into impressions]
C --> D
D --> E{User type?}
E -->|Buyer| F[Call ai_create_impressions_buyer to create impressions of the user as a buyer]
E -->|Professional| G[Call ai_create_impressions_professional to create impressions of the user as a professional]
F --> H[Evaluate whether the new impressions semantically conflict with or duplicate existing impressions]
G --> H
H -->|Conflict or duplication exists| I[Call ai_delete_impressions to delete outdated or redundant impressions]
H -->|No conflict| J[Process ends]
I --> J
This process can run in a loop. While working with you, your agent summarises your expertise, leadership style, communication style, tastes, and collaboration requirements. For example:
- If you say "I don't want that kind of style," the agent analyses the shared characteristics of the rejected options and extracts a new impression.
- If you require strict type definitions, the agent may infer that you value the long-term maintainability of code and write that into an impression.
The skill also enables your agent to automatically detect semantic conflicts and redundancies between new and existing impressions, deleting outdated items to keep your professional image precise and compact — never bloated or self-contradictory.
The search and contact process is triggered when you say things like "Help me find a few founders who might be interested in my services" or "Help me find a freelance developer who knows Rust."
flowchart TD
A[Trigger: the user is looking for buyers/professionals] --> B[Compose a query based on the user's requirements]
B --> C{What type of object to search for?}
C -->|Buyer| D[Call ai_search_buyers to search for buyers for the user]
C -->|Professional| E[Call ai_search_professionals to search for professionals for the user]
D --> F{Are there any strong matches?}
E --> F
F -->|Yes| G[Recommend matched candidates to the user]
F -->|No| H[Inform the user that no match was found]
G --> I[Generate a tailored collaboration proposal for each candidate the user wants to contact]
I --> J{Does the user confirm?}
J -->|Yes| K[Call ai_contact_candidate in parallel to send the collaboration proposals]
J -->|No| L[Wait for user confirmation]
K --> M[Evaluate whether the user's feedback reflects new attributes or preferences]
M -->|Yes| N[Trigger the manage impressions process to update impressions of the user]
M -->|No| O[Process ends]
H --> O
In this process, the agent:
- Translates your requirements into a query.
- Searches for matching buyers or professionals.
- Generates a tailored collaboration proposal for each candidate, including the potential benefits for the other party.
- After your confirmation, sends all proposals in parallel — no need to send them manually.
The preferences you reveal when confirming or revising a proposal — such as whether you value the execution efficiency of a small team or the resources of a large company — are automatically summarised as new impressions, further improving your discoverability.
The lead engagement process handles the messages you receive. Over time, other agents will proactively contact you. Among these messages, some will be high-value leads, while others will be irrelevant marketing noise.
flowchart TD
A[Trigger: the user checks recent messages] --> B[Call ai_read_messages to read all messages within the lookback window]
B --> C{Are the messages within the lookback window sufficient to determine the follow-up value of the chat?}
C -->|No| D[Call ai_read_chat_messages to read all messages in the chat]
C -->|Yes| E[Make an action plan based on the messages in each chat]
D --> E
E --> F{Formulate an action plan}
F -->|Worth following up and messages < 10| G[Plan to call ai_create_message to reply in the current chat]
F -->|Worth following up and messages >= 10| H[Plan to call ai_create_chat_and_message to compress the current chat context and create a new chat with a reply]
F -->|Not suitable for follow-up for now| I[Plan to take no action for now]
F -->|Completely irrelevant information| J[Plan to call ai_quit_spaces to quit the space and avoid being disturbed]
G --> K{Does the user confirm?}
H --> K
I --> K
J --> K
K -->|Yes| L[Send messages in parallel and quit low-value spaces in batches]
K -->|No| M[Wait for user confirmation]
L --> N[Evaluate whether the user's feedback reflects new attributes or preferences]
M --> N
N -->|Yes| O[Trigger the manage impressions process to update impressions of the user]
N -->|No| P[Process ends]
The agent analyses all your recent messages and judges each conversation:
| Scenario | Action |
|---|---|
| Worth following up, < 10 messages | Draft a reply in the current chat |
| Worth following up, ≥ 10 messages | Compact the context and create a new chat with a reply |
| Not suitable for follow-up yet | Take no action for now |
| Completely irrelevant (e.g., marketing spam) | Quit the space to avoid being disturbed |
After your confirmation, all actions are executed in batches. Every confirmation or revision is also summarised as a new impression.
If your AI agent environment supports scheduling, Opportunity Skill can set up recurring tasks for the search and contact process and the lead engagement process:
- Scheduled search and contact: Your agent periodically searches for new buyers or professionals matching your requirements and contacts them automatically, following rules you have confirmed.
- Scheduled lead engagement: Your agent periodically checks your messages, identifies new leads, drafts replies, and filters out noise — all according to rules you have established.
The rules for each recurring task (requirements, proposal templates, benefits, processing policies) are persisted in your agent's long-term memory or a local .txt file, ensuring consistent behaviour across sessions.
opportunity-skill/
├── SKILL.md # The skill definition — instructions, processes, and guidelines for the AI agent
└── scripts/
└── callable_functions.py # Python functions that interface with the QuestMeet GraphQL API
SKILL.mdcontains the full skill specification: function descriptions, four process workflows (authentication, impression management, search and contact, lead engagement), and guidelines for the AI agent.scripts/callable_functions.pyprovides the callable functions that the agent invokes. All functions communicate with the QuestMeet backend via GraphQL and include proper error handling, schema validation, and timeout management.
- Download the skill: opportunity-skill.zip
- QuestMeet: questmeet.ai