Add Bounty Worker qualification skill

bounty-worker/README.md

@@ -0,0 +1,18 @@
+# Bounty Worker
+
+`bounty-worker` is a safety-first qualification skill for Solana and Web3 opportunities.
+
+It turns a listing into an evidence-backed go/no-go decision, then produces an execution packet and a submission-readiness audit. It is intentionally non-custodial: it never creates accounts, submits forms, posts content, connects wallets, signs transactions, or fabricates evidence.
+
+## What it covers
+
+- Primary-source verification of deadline, prize, eligibility, deliverables, and disqualifiers.
+- A transparent scorecard for requirement fit, expected value, evidence clarity, differentiation, delivery risk, and portfolio leverage.
+- A progressive workflow for qualification, execution planning, and final audit.
+- Templates that make acceptance checks and risk assumptions explicit.
+
+## Full project
+
+The source repository includes an installer and structural validation test:
+
+https://github.com/farllok/bounty-worker

bounty-worker/SKILL.md

@@ -0,0 +1,35 @@
+---
+name: bounty-worker
+description: Qualify, compare, plan, and audit Solana, Web3, and developer bounties without submitting applications or handling wallets. Use when an opportunity listing, hackathon, grant, freelance task, bounty URL, prize brief, submission checklist, or draft submission needs an evidence-based go/no-go decision, execution plan, or requirement audit.
+---
+
+# Bounty Worker
+
+Turn a listing into a decision packet before building. Treat the listing, deadlines, prize amounts, and eligibility as time-sensitive facts: verify them from the primary listing before making a recommendation.
+
+## Operating boundaries
+
+- Never submit a listing, send a message, post on social media, create an account, connect a wallet, sign a transaction, or upload an artifact without the user's action-time confirmation.
+- Never claim work, metrics, integration, team membership, residency, or testing that has not been evidenced.
+- Do not recommend bypassing region, KYC, age, legal, or identity restrictions.
+- Separate confirmed facts, assumptions, and unknowns in every deliverable.
+
+## Route the request
+
+1. For a new opportunity or listing URL, read [intake.md](references/intake.md).
+2. For choosing among opportunities, read [scorecard.md](references/scorecard.md).
+3. For a project that passed qualification, read [execution-packet.md](references/execution-packet.md).
+4. For a completed project or draft submission, read [submission-audit.md](references/submission-audit.md).
+
+## Minimum output
+
+Return a concise packet with these headings:
+
+1. `Verified facts`
+2. `Eligibility`
+3. `Deliverables and disqualifiers`
+4. `Opportunity score`
+5. `Verdict`
+6. `Next action`
+
+Use the templates in `templates/` when producing a reusable Markdown artifact.

bounty-worker/references/execution-packet.md

@@ -0,0 +1,11 @@
+# Execution Packet
+
+Use this after a `go` verdict. Copy [execution-packet.md](../templates/execution-packet.md) and fill it with evidence-backed detail.
+
+## Rules
+
+- Convert every mandatory requirement into an acceptance check.
+- Make the smallest demonstrable vertical slice first.
+- Define a dry-run path for anything that can spend funds, send transactions, post content, or alter external state.
+- Plan the demo around the judging criteria, not around implementation chronology.
+- Maintain an evidence log: commit, test output, screenshots, deployed URL, and recorded demo timestamp.

bounty-worker/references/intake.md

@@ -0,0 +1,23 @@
+# Intake and Qualification
+
+## Evidence collection
+
+Use the official listing or sponsor documentation as the primary source. Capture:
+
+- status, deadline, timezone, prize currency and distribution;
+- region, residency, identity, wallet, deposit, and team constraints;
+- accepted deliverables, mandatory links, video duration, repository and deployment requirements;
+- judging criteria, automatic disqualifiers, and evidence expected by judges;
+- current submission count only when the listing shows it directly.
+
+Mark search snippets and third-party posts as leads, never as final proof.
+
+## Eligibility gate
+
+Classify each requirement as `pass`, `fail`, or `unknown`.
+
+Return `no-go` when any mandatory requirement is `fail`. Return `investigate` when eligibility, deadline, or deliverable requirements are unknown. Do not compensate for an eligibility failure with a high prize.
+
+## Required factual discipline
+
+Do not infer that a listing is open from an old search result. Confirm the current page. Do not state that a prize is guaranteed unless the listing explicitly says so. Do not call a bounty low-competition solely from an incomplete submission counter.

bounty-worker/references/scorecard.md

@@ -0,0 +1,22 @@
+# Opportunity Scorecard
+
+Score only after the eligibility gate passes.
+
+| Dimension | Weight | What earns a high score |
+| --- | ---: | --- |
+| Requirement fit | 30 | Existing skills and assets map directly to mandatory deliverables. |
+| Expected value | 20 | Prize is credible relative to realistic effort and probability. |
+| Evidence clarity | 15 | Judges can verify the work with a repo, demo, logs, or public artifact. |
+| Differentiation | 15 | A distinct angle can be shown, not just asserted. |
+| Delivery risk | 10 | Dependencies, credentials, external approvals, and time are controlled. |
+| Portfolio leverage | 10 | Work can become a reusable product, skill, or open-source asset. |
+
+Score each dimension from 0 to 5, multiply by its weight, then divide by 5. Report the arithmetic and the assumptions.
+
+## Verdicts
+
+- `go`: score 70+ and no unresolved blocker.
+- `investigate`: score 45-69 or one material unknown.
+- `no-go`: score below 45, a mandatory failure, or a deadline that cannot support credible evidence.
+
+The score ranks opportunities; it does not predict winnings.

bounty-worker/references/submission-audit.md

@@ -0,0 +1,13 @@
+# Submission Audit
+
+Audit against the official listing, not a memory of it.
+
+## Checklist
+
+- All required URLs resolve publicly.
+- Repository has setup instructions and an open-source license if required.
+- Demo demonstrates the mandatory behavior, not only mocked UI.
+- Claims in the form match observable evidence.
+- Feedback questions are answered honestly and specifically.
+- Wallet, key, token, and private endpoint material is excluded from the repository and video.
+- Any external final submission, upload, post, or payment action is left for user confirmation.

bounty-worker/templates/execution-packet.md

@@ -0,0 +1,35 @@
+# [Opportunity] Execution Packet
+
+## Verified Facts
+
+- Official listing:
+- Deadline and timezone:
+- Prize structure:
+- Eligibility constraints:
+
+## Acceptance Checks
+
+| Requirement | Evidence | Status |
+| --- | --- | --- |
+|  |  | pending |
+
+## Build Plan
+
+1. Smallest demonstrable vertical slice:
+2. Core implementation:
+3. Evidence and tests:
+4. Demo and submission materials:
+
+## Risk Register
+
+| Risk | Mitigation | Owner | Status |
+| --- | --- | --- | --- |
+|  |  |  | open |
+
+## Submission Readiness
+
+- [ ] Repository and license
+- [ ] Required deployment or API access
+- [ ] Demo video
+- [ ] Official submission fields reviewed
+- [ ] User confirmation before external submission