Skip to content

Crowd9/giveaway-skills / skills/giveaway-idea-generator

giveaway-idea-generator

Verified

Generate or score giveaway concepts. Use for 'giveaway ideas', 'Instagram giveaway ideas', 'ideas for my Shopify store', 'win your cart', 'Christmas giveaway', 'ideas for our launch', '10k follower milestone', 'collaboration giveaway', or 'is my idea any good'. Recommend three concepts with hooks, mechanics and Prize directions, or evaluate the user's concept.

FileKindSize
SKILL.mdskill md13.6 KB
evals/cases.mdmarkdown5.7 KB
evals/evals.jsonconfig18.5 KB
references/evidence-detail.mdmarkdown3.2 KB
references/hook-patterns.jsonconfig2.0 KB
references/hooks-and-themes.mdmarkdown46.3 KB
references/house-style.mdmarkdown4.7 KB
scripts/style_check.pyexecutablescript40.3 KB

references/house-style.md

View on GitHub
<!-- Generated from scripts/house-style.md by scripts/render_shared.py. Edit the source, never this copy. -->

# House Style Detail

## Figures and Comparisons

Match the wording to what the source counts. Shares of distinct people or campaigns are percentages: "48% of Entrants followed on Instagram" only when the source counts distinct followers, "referred Entrants were 19% of the total" when it counts distinct referred Entrants. Write completed events as a percentage of the Entrant count: "0.52 joins per Entrant" becomes "join completions equal to 52% of Entrants". Repeats count, so 52 join completions from 26 of 100 Entrants means join completions equal to 52% of Entrants. The distinct-person count supports "26% of Entrants joined". Only a source counting distinct people supports "52% of Entrants joined". Counts above one per person may use "each": "2.5 Entries each", "1.5 visits each". Comparisons use the letter x with no space: "1.5x the typical campaign". Reference tables keep per-100 columns, which can exceed 100. Translate those cells to percentages or "each" in the answer, naming completions when the source counts events.

Quote at most two figures per point, chosen because they decide the reader's question. Leave other figures in the reference, including figures that would only lengthen a table. If the reader's size is unknown, ask or identify the middle case you use.

Copy every figure from the open reference. Before sending, search the reference for every quoted figure, then check its row label and column header against the sentence. Check Entry Method counts against what was counted inside an Action, and Prize-value-per-email against stated Prize value. A configuration row supports only the dimensions it contains.

Use the count printed beside the figure's own table. A table headed "1,000 or more Entrants" describes that group. A rate per Entrant measures a different thing from a share of campaigns. "Email signups per Entrant, where offered" at 0.96 means campaigns offering it recorded signup completions equal to 96% of Entrants. Describe a share of people only when the source explicitly counts distinct people.

Share one sample count only across metrics with the same population. Otherwise keep only metrics whose sample counts and relevant missing-data notes fit one short source line. Give that line only when quoting figures, in plain words naming the campaigns counted. Keep file names, keys and "extracted" internal. Leave source blocks and extra figures in the reference.

Take both sides of a comparison from the same table and use its like-for-like row. Read both cells before claiming one is higher. Recompute every stated difference from the figures printed beside it: 403 against 460 is about 12% fewer. Retain only differences that the quoted figures reproduce. Before a superlative such as "most", "largest", "best" or "cheapest", sort and check the relevant column.

For example, describe a value index of 0.99 as "drew about as many Entrants as campaigns offering a Prize of similar stated value".

## Claims and Tradeoffs

Review each sentence about how a channel, platform, service or person behaves, including every `unsourced_claims` flag. Name its basis in that sentence, rewrite it as an action, or remove it. Use wording already supported by the reference. For example, "Screenshot the account before you report it, and post your official handle so Entrants can tell the difference" gives a useful action. Product specifications, base rates and algorithm behaviour need a source to remain. Judge support for the claim itself, regardless of whether rewording clears the checker.

Reserve categorical warnings for platform rules, legal risks and risks to the business's money or data, subject to the legal guardrail in the evidence core. For other choices, state the tradeoff: what it costs, what it buys and when it fits.

## Style Verdicts and Manual Checking

When the reader requests a style verdict, lead with the verdict in a sentence, identify any failed counter and the change needed, then paste the script's output file inside a fenced block. Copy counters from that file.

Without a shell, do a manual last pass: search for ", not ", "rather than", "instead of" and "not X but", and rewrite contrast-led sentences as direct statements. Cut "actually", "leverage", "robust", "comprehensive", "streamline", "delve", "furthermore", "moreover" and "it is worth noting". Write colon reveals as sentences. For answers with at least 40 words of prose, include one sentence of eight words or fewer and one of eighteen words or more, and end with a concrete detail, instruction or question. Shorter answers can give the requested fact or copy without padding for sentence variety or the ending check. Code blocks, URLs and tables do not count toward this threshold.
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 196kUpdated Claude CodeCodexVerifiedInstall