Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/xxyoudeadpunkxx/signal-rail/llms.txt

Use this file to discover all available pages before exploring further.

Guided prompts are structured, step-by-step question flows stored in 07_guided_prompts_test.txt. They exist for higher-risk operating situations — moments where a normal session opening is not enough to safely close the working frame before action. Each flow reduces ambiguity one question at a time, in a specific order, before the agent reads, routes, rewrites, or rebuilds state. They are not casual prompt examples and are not a replacement for the full operating protocol.

​
Common Rules

These rules apply to every guided prompt flow, regardless of which path you use:
  • Read 00_runtime_entry.txt first, then read and follow 06_ai_to_ai.txt
  • Ask one question at a time — never dump multiple intake questions at once
  • Allow free-form answers; reduce ambiguity before moving forward
  • Do not move on until the current point is clear enough
  • Refine the same point if the answer is partial
  • Skip what is already clear
  • Use the user’s language unless asked otherwise

​
The Seven Guided Flows

Goal: Open a safe working session in Signal Rail. Close the minimum sufficient frame — host project, authority, working object, mode, session goal, and source scope — before any substantive action.When to use: At the start of a session inside a project that is already live enough to work on.Questions asked, in order:
  1. What is the host project? (Closes the real project before any mode or container decision.)
  2. Who has decision or document authority on this project? (Tells you who can confirm freeze, live state, and source conflicts.)
  3. What is the current working object? (Separates project work, source analysis, and system work.)
  4. What is the active mode for this session? (Limits allowed actions before reading or writing.)
  5. What is the current session goal? (Prevents scope drift.)
  6. Which sources were explicitly indicated for this session? (Tells you what may be read without guessing.)
  7. Is any live source or authority still uncertain? (Tells you whether to remain in reference or review.)
Return format:
host project:
owner or authority:
working object:
mode used:
sources to read next:
what remains uncertain:
recommended next step:
Goal: Open a new or near-blank Signal Rail instance without inventing project reality. Close real use, perimeter, authority, and first live objective before writing.When to use: When a Signal Rail instance is new, almost empty, or still mostly template — no reliable 01–03 yet.Questions asked, in order:
  1. What is the host project? (Closes the real project that Signal Rail will govern.)
  2. Who validates project truth? (Tells you who can confirm what is real versus template, draft, or idea.)
  3. What is the project, in one clear description? (Prevents filling orientation from assumption.)
  4. What is Signal Rail supposed to govern here? (Separates governance from product, implementation, and support material.)
  5. What is already real, and what is still template, sketch, or idea? (Prevents fake bootstrap.)
  6. What is the current project perimeter? (Defines what is in scope now and what stays outside.)
  7. Where does the project really live technically right now? (Closes the first version of 08 without guessing.)
  8. What is the first real operational objective? (Gives 03 a real present state instead of filler.)
Return format:
host project:
authority:
mode used:
minimum sources to read:
what is already real:
what is still template:
technical surface to map:
what remains uncertain:
recommended next step:
Goal: Receive a raw idea in human language and place it at the right level of the project without promoting it too early.When to use: When the user says “I have an idea,” “maybe we should,” or offers a new thought without structure.Questions asked, in order:
  1. What is the idea, in your own words? (Preserves original meaning before classification.)
  2. Why does this idea matter now? (Separates live leverage from loose speculation.)
  3. Does this feel more like a deep direction, a next move, or something to keep alive for later? (Narrows the right level without forcing internal labels too early.)
  4. Is this idea already guiding current live work, or is it still only a proposal? (Prevents premature placement into 03.)
  5. What is still uncertain about it? (Tells you whether it can move now or should stay unresolved.)
  6. Does it already have the shape and stability of live work or a decision, or should it stay mobile for now? (Prevents early crossing into 03 or 04 when the idea still belongs in 05.)
Return format:
host project:
working object:
mode used:
extracted idea:
likely place:
local 05 id if needed: use a new L-xx; keep any D-xx only as reference
external reference if needed: use external reference, not links to
what remains uncertain:
recommended next step:
Goal: Handle messy, scattered, or misaligned sources before safe ingest or rewrite. Rebuild source authority and current live state before promoting, rewriting, or classifying.When to use: When many files are in scope, the sequence is unclear, authority is uncertain, or canonical files are out of sync.Hard boundary: Stay in reference or review until authority is clear enough for safe ingest or rewrite.Questions asked, in order:
  1. What is the real host project for this material? (Closes the project before reading mixed material.)
  2. What is the current working object? (Separates project work, source analysis, and system work.)
  3. Which sources are intentionally in scope for this pass? (Limits reading to declared material.)
  4. What kind of confusion is present here? (Distinguishes scattered sources, timeline uncertainty, and canonicals out of sync.)
  5. Which sources look most authoritative, and why? (Surfaces the governing candidates before classification.)
  6. Which sources best reflect what is live right now? (Separates current state from old truth or cleaner drafts.)
  7. Which sources look newer, cleaner, or more complete but may still be less authoritative? (Blocks false promotion from appearance.)
  8. What do you want from this pass: authority map, timeline rebuild, conflict map, or proposed destinations? (Closes the minimum safe output.)
  9. What is the minimum useful output of this pass? (Prevents overreach.)
  10. What is the smallest unresolved hinge still blocking safe continuation? (Keeps the pass narrow when authority is not fully closed.)
Return format:
host project:
working object:
mode used:
candidate sources:
presumed source roles:
authority conflicts:
timeline anchors:
current live state:
safest next move:
what remains uncertain:
recommended next step:
Goal: Review authority, risks, and rewrite safety before changing text. Separate what is truly governed or decided from what only sounds settled. Do not rewrite until the governing source is clear enough.When to use: Before rewriting a file, section, or source set that may be cleaner than it is true.Hard boundary: Stay in review mode throughout.Questions asked, in order:
  1. What text or set of texts is under review? (Closes the rewrite target.)
  2. Which sources may govern that target? (Identifies live authority before cleanup.)
  3. What seems already governed or decided, and what still sounds open? (Prevents rewriting open material as if it were settled.)
  4. Where do you suspect conflict or distortion risk? (Surfaces why rewrite may be dangerous.)
  5. What outcome do you want from review? (Prevents premature rewriting.)
Return format:
host project:
working object:
mode used:
governing sources:
likely settled area:
likely open area:
conflicts found:
rewrite-safe area:
rewrite-unsafe area:
recommended next step:
Goal: Place one extracted unit in the right destination without routing a whole file. Decide whether the unit stays live, waits, or becomes reference only.When to use: When one unit clearly matters, but its right destination is still uncertain.Questions asked, in order:
  1. What exact unit are we placing? (Keeps the real unit small enough to classify safely.)
  2. Where does the unit come from? (Preserves source context and authority clues.)
  3. Is this unit project identity, a stable constant, current live work, a taken decision, technical surface map, useful but inactive, closed history, or still unresolved? (Closes level before destination.)
  4. What shape and stability does this unit already have? (Reduces routing anxiety and helps place the unit only after its level is clearer.)
  5. Why does the unit still matter? (Distinguishes live unresolved material from dead history.)
  6. What still prevents placing it safely now? (Tells you whether it should stay live, wait, or become only reference.)
Return format:
host project:
working object:
mode used:
source:
extracted unit:
proposed destination:
what remains uncertain:
recommended next step:
Goal: Reconstruct a defensible project timeline before writing canonicals. Separate historical reconstruction from canonical promotion. Backfill Signal Rail for a project that was not governed from day one.When to use: When project history is short enough to reconstruct but dense enough to create level confusion — or when the current session holds a lot of memory but the canonical files were created late.Hard boundary: Build the timeline first. Do not write canonicals first. Do not invent missing history to complete the sequence. If a phase, transition, or destination is still unclear, mark it uncertain instead of forcing promotion.Questions asked, in order:
  1. What is the host project? (Closes what history is being reconstructed.)
  2. Which sources are valid for reconstructing the project history? (Closes the defensible source set before interpretation.)
  3. What are the major phases, versions, or turning points of the project? (Creates the first timeline skeleton.)
  4. For each phase, what changed, why did it matter, and what did it produce? (Separates real turning points from noise.)
  5. For each phase, what is its status today? (Distinguishes live, won, latent, parked, and dead material.)
  6. Which timeline units belong to 01, 02, 03, 04, 05, 08, 98, or 99? (Turns raw history into level-aware canonical candidates.)
  7. Which units are still too uncertain to promote safely? (Prevents fake closure.)
  8. Do you want timeline only, timeline plus destination map, or timeline plus canonical draft proposals? (Closes the safe output scope before writing.)
Return format:
host project:
mode used:
source set:
reconstructed timeline:
likely canonical destinations:
what remains uncertain:
recommended next step:
When unsure which prompt fits best, describe the situation to your AI agent and ask it to recommend the best-fit guided path from 07_guided_prompts_test.txt.

Build docs developers (and LLMs) love

Get started for freeTalk to us