Executive Upward Communication Skill
Overview
Given raw, often poorly-framed user text describing a request or proposal intended for a more-senior stakeholder, the agent must produce a crisp, decision-ready upward-communication package — a structured one-pager that leads with the recommendation, supplies evidence, surfaces risk, and closes with an explicit decision line.
Your submission is a single SKILL.md file that teaches a fixed base agent how to receive messy input and convert it into a correctly structured executive communication — calibrated to the recipient's seniority, free of background narrative, and actionable on first read.
Strong skills teach the agent a transferable procedure: classify the request type, invert the default context-first instinct, apply a fixed four-part structure, calibrate density to the recipient tier, and flag gaps or unsupported claims rather than papering over them. Weak skills hand the agent a fill-in-the-blank template — and break the moment the request type, seniority level, or missing-data pattern changes.
The skill must also teach the agent to handle judgment calls that arise before drafting even begins: recognising when the raw input bundles multiple request types, when a classification is ambiguous enough to warrant a clarifying question, and when numeric claims need a bracketed verification note rather than confident-sounding laundering.
A skill built to memorise specific scenarios will not hold up as challenges evolve.
You submit a SKILL.md file. A fixed base agent loads it and answers all 10 scenarios with it, then returns a scored breakdown. The submission that holds the top of the leaderboard accrues on-chain emissions for as long as it stays first.
We are looking for skills that demonstrate good design and help the agent perform like a capable practitioner in this domain.
Requirements (10 Tests)
Your skill is tested across four capability types, with 10 scenarios in total. Every submission is evaluated against all 10 scenarios — there is no sampling, so consistent quality across every capability type counts.
| Category | Overview | Test cases |
|---|---|---|
| Type A — Access, Tooling & Vendor Approval | Tests whether the agent can produce a decision-framed subject line that names the specific approval being sought — not merely the topic — so a recipient could forward it with 'approve?' as a complete instruction. Also tests whether the agent identifies and explicitly flags missing critical elements (such as funding source, rollback plan, or evaluation criteria) rather than generating a polished-sounding package that silently conceals gaps in the raw input. | 1–2 |
| Type B — Budget Ask | Tests whether the agent inverts the natural context-first impulse and leads immediately with the recommendation and the decision required, rather than burying both after paragraphs of background. Scenarios in this type present raw input that, if reproduced naively, would place the actual ask at the end — and the agent must restructure it so the decision need is visible within the first two sentences. | 1–2 |
| Type C — Politically Sensitive Request | Tests whether the agent correctly classifies a request before structuring its response, including recognising when a request straddles multiple types (such as a policy exception that also carries headcount or budget implications) or when the classification is ambiguous enough to warrant a clarifying question. The skill must teach the agent to name the request type explicitly and to surface ambiguity rather than resolve it silently. | 1–2 |
| Type D — Strategic Recommendation | Tests whether the agent handles unsupported or vague numeric claims responsibly — inserting bracketed user-to-verify notes rather than laundering uncertain figures into confident-sounding evidence. Scenarios in this type include inputs with plausible but unverified projections, estimates, or benchmarks that a naive agent would reproduce as established fact. | 1–2 |
Surface details — names, titles, dollar amounts, vendor names, timelines, and supporting figures — rotate across scenario variants, while the underlying capability being tested remains the same.
Datasets
Available Datasets
When you submit, we run your SKILL.md through a standardised base agent provided by sundae_bar. The base agent is the same for every submission and is not supplied by you. It receives the challenge's test scenarios one at a time, calls the load_skill tool to read your skill, and follows it to answer each scenario.
Same base agent, same tests, only the submitted skill changes. Scores therefore reflect the quality of your skill, not agent configuration: the agent's answers are only as good as the procedure your skill teaches.
Graders then score each answer for output quality, correct skill use, and gaming behavior, and the per-scenario scores are averaged into your challenge score.
Important: the base-agent harness is fixed at evaluation time and its available tools are load_skill. You submit only a SKILL.md; you cannot supply your own agent, add tools, or alter the base agent's configuration.
Tip: write your skill as though you are onboarding a capable new hire who needs a procedure, not scripting a chatbot.
Evaluation Criteria
Each scenario is graded by an LLM judge. Final scores combine two graders and a set of deterministic penalties.
| Criteria | Weight | |
|---|---|---|
Skill use An LLM judge evaluates whether the agent's output actually followed the procedures your skill teaches, rather than freelancing or pretending to use the skill. | 50 % | |
Scenario quality An LLM judge evaluates how well each output meets the per-type quality bar, independent of whether your skill was followed. A reference good output anchors the judge's sense of the quality bar. | 50 % |
Scenarios are scored individually; deterministic penalties apply when the agent's output contains moves that materially damage quality — such as burying the recommendation, omitting the decision line, fabricating or laundering unverified claims, or silently papering over missing critical elements. Individual scenario scores aggregate into per-capability-type averages, which in turn roll up into the final challenge score shown on the leaderboard.
Submission
How to submit
Create an account and add your Bittensor wallet ID to your profile to earn alpha rewards in the challenge.
Review the reference skills (good, mediocre, gaming) to understand the quality bar and calibration points.
Draft your SKILL.md. Focus on teaching procedure rather than providing templates.
Validate locally using the published dataset variants to sanity-check output quality.
Submit through the sundae_bar challenge portal by uploading your SKILL.md file.
Review your scored breakdown and resubmit.
Constraints
Submission format: a single SKILL.md file with YAML frontmatter (name, description) and Markdown body.
Base agent: fixed. Provided by sundae_bar.
Runtime: a fixed base-agent harness loads your SKILL.md and answers all 10 scenarios per submission with it.
Tests per submission: 10 (the full scenario set — no sampling).
Submission limits: 3 submissions per day. A rolling challenge limit also applies.
Rewards
Emissions accrue on-chain to whichever submission currently holds the top of the leaderboard. There is no fixed per-submission payout: the current leader earns for as long as it stays first.
Time at the top takes all. A strong skill submitted early starts accruing immediately and keeps earning until another submission displaces it. Once you are displaced, accrual stops until you reclaim the top.
Rankings shift as others iterate. Get to the top early, then keep refining your skill to defend the position.
This challenge miners earn 40% of emissions.
Your SKILL.md is scanned at submission time by a skill integrity gate. The gate checks that your skill teaches procedure rather than encoding scenario-specific content from our evaluation pool.
A skill that fails the integrity gate scores 0 for the submission. The gate runs deterministically and reports its finding as part of your results. This is how the evaluation framework stays credible: a high score should reflect a skill that genuinely teaches procedure, not one that memorised our test set.