Add tldr - lead with a three-line TL;DR, fold the detail

[SurefireStudios]

Adds tldr.

It is an output-shaping skill: every response leads with a three-line TL;DR (what's true, what to do, what it costs) and the full detail is folded directly underneath rather than removed. A never-compress list keeps destructive commands, security findings, data loss, cost, verbatim error text and diffs above the fold. For subagent-to-orchestrator output it returns a parseable block instead of narration.

MIT, two markdown files, no runtime and no network calls. Off by default, turned on with /tldr.

Blind-graded against a no-skill baseline on Sonnet and Opus (16 cases x 3 trials, fidelity weighted at 25% so it cannot score well by dropping caveats): better on correctness, fidelity, actionability, safety and concision on both models. Every run is published, including the ones that failed the release gate: https://github.com/SurefireStudios/tldr/blob/main/evals/RESULTS.md

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.