Market Data Comparability Assessment Skill
Overview
This skill teaches the agent to evaluate retrieved market and competitor figures and decide whether each one is safe to place next to an internal company metric in a board narrative, pricing memo, or executive briefing.
Given a pool of retrieved figures, drawn from structured data rows, filings, analyst reports, press releases, transcripts, and other sources, alongside the internal KPI each figure would sit next to, the agent produces a structured comparability table that assigns every figure a confidence tier (HIGH, MEDIUM, LOW, or REJECT) and a one-line rationale grounded in the specific evidence present in that scenario.
The submission is a SKILL.md, with optional reference files the agent opens by name when SKILL.md directs it to. Strong skills teach a reasoning procedure the agent can apply to any retrieved figure: how to check each equivalence dimension, how to weigh evidence quality by source type, and how to write rationales that name actual scenario evidence rather than restate a generic rule. Weak skills that supply templates or pre-filled answers break the moment surface details change.
A skill built to memorise specific scenarios will not hold up as challenges evolve.
You submit a skill: a SKILL.md and, if you need them, reference files it points at. A fixed base agent loads it and answers all 9 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 (9 Tests)
Your skill is tested across 5 capability types, with 9 scenarios in total. Every submission is evaluated against all 9 scenarios, there is no sampling, so consistent quality across every capability type counts.
| Category | Overview | Test cases |
|---|---|---|
| Type A: Accounting Basis & Period Headline Rejection | This type tests whether the agent correctly identifies and rejects figures that carry confirmed accounting-basis or period mismatches relative to the internal KPI, even when the figure is otherwise well-sourced or numerically plausible. The skill must teach the agent to apply the strict downward-pressure tier rule, assigning REJECT when a mismatch is confirmed, and never softening to LOW or MEDIUM simply because some dimensions appear compatible. | 1 |
| Type B: Definition Mismatch with Non-Rescuing Footnote | This type tests whether the agent correctly distinguishes a confirmed definition mismatch from an unconfirmed one when a quantified footnote is present but does not resolve the incompatibility. The skill must teach the agent to read footnote content carefully, recognise when a footnote narrows but does not close a definitional gap, and assign REJECT only where the mismatch is confirmed, preserving LOW for figures where cohort inclusion or a reasonable inference remains available. | 1 |
| Type C: Denominator Methodology Mismatch with Strategic Narrative Implication | This type tests whether the agent surfaces silent assumptions about denominator methodology when the retrieved figure's calculation basis differs from the internal KPI's, and when the directional implication of that difference matters to the narrative's intended use. The skill must teach the agent to name the assumption explicitly in the rationale rather than resolving the ambiguity by treating the figures as equivalent without disclosure. | 1 |
| Type D: Denominator Methodology Mismatch Without Explicit Blocker Signal | This type tests whether the agent identifies denominator methodology mismatches even when no explicit signal (such as a contradicting footnote or stated exclusion) is present to flag the problem. The skill must teach the agent to reason from what is absent, the lack of a stated methodology, and to surface that silence as an assumption in the rationale rather than defaulting to equivalence. | 1 |
| Type E: Multi-Figure Definition and Segment Mismatch | This type tests whether the agent produces rationales that are specific to each individual figure in a multi-figure scenario, naming the particular definition or segment evidence that drives each tier assignment. The skill must teach the agent to avoid rationale language that could be copy-pasted across rows, ensuring every one-line rationale reflects the planted evidence unique to that figure. | 1 |
Surface details such as company names, metric labels, reported values, source names, and fiscal periods rotate across scenario variants while the underlying comparability reasoning challenge remains the same.
Datasets
Available Datasets
When you submit, we run your skill 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.md and the list of your reference files, opens a reference with read_reference when your SKILL.md tells it to, and follows the skill 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.
What you submit: a folder or a .zip with SKILL.md at the top level and an optional references/ folder. A single SKILL.md is also accepted. All files are the submission.
Limits: SKILL.md is at most 10,000 bytes and is always loaded. references/ holds up to 30 files, each at most 32,000 characters, 128,000 characters in total. Text formats only: .md, .txt, .json, .yaml, .yml and .csv. Lower-case file names using letters, digits, dots, dashes and underscores. The current values are published by the platform API at GET /api/v2/submissions/limits; the submission form reads them from there.
How references are read: the agent always sees SKILL.md plus a list of your reference file names. It opens a reference only when SKILL.md tells it to, by name. Each read is one harness step within a cap of 10 steps per scenario, and the file's text counts against the scenario's token budget of 37,000 tokens. A reference the agent never opens does nothing for the score.
Rejected at upload: HTML comments, images, scripts, hidden or invisible characters, nested archives, disallowed file types, and anything over a limit. Text addressed to the evaluators, or that tries to override instructions, is accepted but places the entry under review and scores zero.
Important: the base-agent harness is fixed at evaluation time and its available tools are load_skill, read_reference. You submit only skill files; 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 four graders and a set of deterministic penalties.
| Criteria | Weight | |
|---|---|---|
Response quality An LLM judge scores each response against a task-specific quality bar for this challenge, how correct, complete, and usable the output is on its own terms, independent of which procedure produced it. | 30 % | |
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. | 30 % | |
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. | 30 % | |
Output originality A check that your skill drives genuine per-scenario reasoning rather than emitting near-identical, templated responses. Skills that reuse one canned output across scenarios lose credit here. | 10 % |
Scenarios are scored against a structured rubric; deterministic penalties apply when the agent's output damages downstream citation safety, for example by passing through a figure with a confirmed mismatch, producing a rationale that does not engage with the scenario's actual evidence, or omitting actionable re-retrieval asks on rejected figures. Individual scenario scores aggregate to capability-type averages, which combine 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 (at most 10,000 bytes; it is always loaded) and, if the method needs them, put tables, templates and worked examples in reference files under references/. Tell the agent in SKILL.md which reference to open and when, by file name. 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 folder, a .zip of it, or a single SKILL.md. The folder has SKILL.md at the top level and an optional references/ folder beside it.
Review your scored breakdown and resubmit.
Constraints
Submission format: a folder or .zip with SKILL.md (YAML frontmatter with name and description, then a Markdown body) at the top level and an optional references/ folder. A single SKILL.md is also accepted.
Limits: SKILL.md at most 10,000 bytes. Up to 30 reference files, each at most 32,000 characters and 128,000 characters in total; text formats only (.md, .txt, .json, .yaml, .yml and .csv); lower-case file names.
Reference reads: the agent opens a reference only when SKILL.md names it. Each read is one of 10 steps per scenario and its text counts against the scenario's 37,000-token budget.
Base agent: fixed. Provided by sundae_bar.
Runtime: a fixed base-agent harness loads your skill and answers all 9 scenarios per submission with it.
Tests per submission: 9 (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.
Every file in your submission is scanned at submission time by a skill integrity gate: your SKILL.md and every reference file, whether or not your skill points at it. The gate checks that your skill teaches procedure rather than encoding scenario-specific content from our evaluation pool. Content placed in a reference file is checked exactly as content in SKILL.md.
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.
Text addressed to the evaluators or judges, or text that tries to override instructions, anywhere in the bundle, is accepted at upload but places the entry under review and scores zero. Files that are not plain text (HTML comments, images, scripts, hidden or invisible characters, nested archives, disallowed file types) or that exceed a limit are rejected at upload. The copy and similarity checks run over the whole bundle, chunk by chunk, so padding a copied skill with extra reference files does not pass them.