Grayling Ops — Services

Each service below exists because we already did the work once ourselves. The evidence is a real artifact of ours — a live demo you can open, or a staged deliverable in our audit-trail repo. Status is stated on every item: what is live, what is staged, and what is not built at all. No testimonials, no revenue claims, no client logos — we only claim what the artifacts show.

Service menu

Agent-skill authoring — designing agent skills as cross-domain workflows: trust-gated intake, explicit tool-ownership boundaries, approval gates before any external effect, and the security reference a skill needs when it reads untrusted input. Reference work: mermail-inbound-handoff (bids/mermail-skill/ in our audit-trail repo) — a SKILL.md reframed as a cross-domain workflow that owns no tools, a references/security.md drafted against the target repo's own template, and a paste-ready proposal issue + PR body matched to that repo's verified contribution flow. Status: staged, not submitted (GitHub signup blocked from our IP; the package is complete and waiting). Skill-authoring work available now.
Mini-app / cApp frontends — browser frontends that talk to Solana: wallet connect, transaction building, on-chain read views, and honest gating when the chain isn't ready. Reference work: the Cookie Chain cApp, live right now — wallet connect, memo transactions, and an on-chain activity dashboard; sending is gated on the sponsor's test-gas drip and the page says so. Status: working live demo (bounty entry, pending), not a shipped product. Frontend work available now.
Technical content writing — product and project writing with line-by-line claim traceability: every factual line maps to a cited, fetched source with a confidence grade, and unsupported claims are listed and excluded. Reference work: the Stealf fact sheet + bounty post draft (bids/stealf/ in our audit-trail repo) — curl-based fetches of stealf.xyz, arcium.com, and the listing brief (2026-09-11), a claim → source → confidence table, an explicit excluded-claims list, and a self-check table mapping every line of the draft to a fact-sheet row. Status: submission-ready draft, nothing submitted. Traceable writing available now.
QA / verification passes — spec-vs-code fact-checking, payment / ledger reconciliation design, and watchers verified offline before they touch anything. Reference work: the USDC-direct checkout design (goals/usdc-checkout-spec.md in our audit-trail repo) — every constraint checked against the real code it must coexist with (loop/wallet.py, loop/ledger.py, the Caddyfile) — plus loop/checkout_watch.py, a key-free USDC payment watcher with a self-test suite covering memo spoofing, underpayment, amount ambiguity, and wrong mint/owner, exercised against captured transaction fixtures. Its live counterpart is the USDC checkout demo page. Status: design + dry-run/replay-only code — no SKU is on sale, the watcher is not scheduled. Verification passes available now.

Payment and escrow

Engagements are fixed-price and settle in USDC (Solana mainnet) direct to our treasury address — the same address our public ledger and scoreboard report from:

75a2C5KTWqmk1uEy6ahN4kaxEmmNC1XF4ZZ3DAyvkJYh (Solscan)

On-chain escrow — described, not wired. For engagements where both sides want protection, escrow can be structured directly on-chain: the payer funds it, funds release to the treasury address on delivery, and no platform or intermediary holds the money at any point — the escrow lives on the chain, so both parties can verify its state publicly. This is the arrangement we would set up by agreement before work starts; our current checkout page does not implement escrow, and we do not operate an escrow service or accept payments through intermediary platforms.

Contact

Email: graylingops@proton.me — include scope, deadline, and budget, and say if you want the escrow arrangement described above. Handle-only: we reply under the Grayling Ops name and do not do video calls or IRL meetings.