Cue deck
Groundledger: MIT agent skills for Claude Code, Cursor & Cowork — grounded blog SEO/AEO, PROBE site SEO, SAFE paid media, SPARK editorial, SNAP social, ORBIT discovery
npx -y skills add dranshrad/groundledger --skill cue-deckAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 12 days oldThe repository was created 12 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Groundledger Cue Deck: original prompt cards keyed to ORBIT stages and CLEAR modes (When, Inputs, Prompt, Output, Risks). Use when the user wants a reusable prompt, strategy cue, or agent instruction for content work.
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
1.7 KB, 409 tokens by cl100k_base, as published. Nobody here has run it
Cue Deck
Groundledger prompt cards. Every card that asks for facts inherits the Grounding Law — a drawn card never licenses an unledgered number.
Modes
| Mode | Job |
|---|---|
list | Show card index |
draw | Emit one card by id or stage |
custom | Build a new card from a user goal |
Card schema
Every card uses:
## Cue — [id]
**When:** …
**Inputs:** …
**Prompt:**
> …
**Output:** …
**Risks:** …
**Next skill:** blog-engine | orbit-discovery | site-signal | paid-cast | social-cast | editorial-pass | studio-desk
Index
Load full text from references/cards.md:
| ID | Stage / mode |
|---|---|
O-jobs | Observe — JTBD spray |
R-echo | Reinforce — Echo Map |
B-brief | Build — asset brief |
I-kpi | Instrument — measurement |
T-retro | Transmit — learnings |
C-draft | CLEAR draft kickoff |
C-verify | Claim Probe+ |
C-cite | Cite Surface |
S-atom | Social atomize |
S-readout | Pulse readout |
E-spark | SPARK kickoff |
D-ship | Ship Gate |
P-audit | PROBE site audit |
P-fit | Page-fit check |
A-snap | Paid Snapshot (observe-only) |
A-latch | MutationLatch checklist |
When drawing, copy the prompt block verbatim into the agent turn, then execute via the Next skill.
Gives 0 of the 12 instructions most slides presentations skills give in 409 tokens
Counted across 568 of the 571 authors here whose files we hold, read 2026-08-06
- include a visual element on every slidein 52 of 568, across 21 files
- put one idea per slidein 50 of 568, across 42 files
- state the design approach before writing codein 38 of 568, across 8 files
- validate XML immediately after each editin 37 of 568, across 7 files
- rasterize gradients and icons as PNG images before referencing themin 37 of 568, across 7 files
- generate and inspect thumbnails to validate layoutin 37 of 568, across 7 files
- commit to a single visual motif across every slidein 37 of 568, across 12 files
- use web-safe fonts onlyin 36 of 568, across 7 files
- keep 0.5 inch minimum marginsin 35 of 568, across 9 files
- use two-column layout for slides with charts or tablesin 34 of 568, across 5 files
- save a template inventory analysis to a filein 33 of 568, across 4 files
- use subagents to visually inspect rendered slidesin 33 of 568, across 9 files
Said here and by no other author read
- apply the grounding law to fact-based cards
- show the card index in list mode
- emit one card by id or stage in draw mode
- build a new card from a user goal in custom mode
- use the specified markdown card schema
- copy the prompt block verbatim into the agent turn
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.