Ridgtof's templates
Jump to navigation
Jump to search
General[edit]
Author Style Extension[edit]
WIP
Secret Info[edit]
Secret Scheme Record in secretInfo[edit]
A surprisingly effective approach (with Lion, at least) for recording secret NPC plots and schemes without consuming too much space or bloating costs overmuch. This works especially well when used with a background plot/scene to facilitate progress beyond the main scene.
<schemeData>
<schemeDiscoveredValues>
This is a list of possible values and their meaning for the Discovered variable in <schemeDataFormat>:
Hidden - The scheme remains hidden from my knowledge and/or the knowledge of those loyal to me
Detected - The scheme is known to me and/or those loyal to me, but the members of the scheme do not know this to be the case
Suspicious - The scheme is known to me and/or those loyal to me, and the members of the scheme think this might be the case
Alerted - The scheme is known to me and/or those loyal to me, and the members of the scheme know this to be the case
</schemeDiscoveredValues>
<schemeStatusValues>
This is a list of possible values and their meaning for the Discovered variable in <schemeDataFormat>:
New - The scheme was just created and only the Instigator and whomever they were in close proximity to is a member, and is still creating a plan. This can only occur once in the lifetime of a scheme.
Planning - The Instigator and members of the scheme are planning how to accomplish the objective of the scheme with the information they have
Progressing - The scheme is in progress, trying to gather information, position members, and orchestrate events to carry out the Current Plan
Executing - The scheme is enacting the Current Plan, something that will potentially take a long time
Complete - The scheme has successfully completed its Objective. The instigator, members, objective, and results of the scheme should be noted in summary. This can only occur once in the lifetime of a scheme, and is exclusive with Foiled.
Foiled - The scheme has failed to complete its Objective and cannot be recovered. The instigator, members, objective, and results of the scheme should be noted in summary. This can only occur once in the lifetime of a scheme, and is exclusive with Complete.
</schemeStatusValues>
Use this format for writing out ongoing schemes/plots in secretInfo:
<schemeDataFormat>
{scheme_n}
Instigator: {string}
Members: {string list}
Objective: {string}
Current Plan: {string}
Discovered: {string}
Status: {string}
<|stop|>
</schemeDataFormat>
Swap the label '{scheme_n}' for a scheme's name. Replace instances of '{string}' with appropriate text for the variable. Replace instances of '{string list}' with appropriate contents for the variable. Each scheme's data should be separated by newlines. You need to print out all the member variables of a scheme's data for each scheme because you need that information for the next turn. It is thus crucial for you to write out scheme variables word-for-word in every single response, including this one. Never simply refer to the previous turn, always write the scheme variables in complete detail. Objective never changes. Instigator never changes. When a scheme's Status is either Complete or Foiled, it no longer needs to be reproduced in secretInfo, and can be discarded once its Instigator, Members, Objective, and Results have been noted in summary.
Never terminate output for a scheme before the stop token, '<|stop|>'. Do not include the stop token in output; delete it from secretInfo.
</schemeData>
...
<secretInfoInstructions>
...
step_n. [...] [U]sing the rules and format outlined in <schemeData>, update and print a description of every secret plot or scheme in progress from either the main scene or the background plot. This information is essential to informing how schemes and events progress, what machinations are in motion, why characters are doing what they do, and what the scheme's members intend to do or are doing. It is vital that you include every scheme in this section.
...
<secretInfoFormat>
...
{scheme_n data}
...
</secretInfoFormat>
Replace every instance of '{scheme_n data}' with suitable, innovative writing and information based on <schemeData>.
</secretInfoInstructions>
Character Tracker[edit]
WIP
Evaluation Management[edit]
Simple Skill Check Override System[edit]
This is a very basic example of how to tell the AI to manage skill checks without necessarily using the vanilla skill system. This can be accompanied by a description of each skill (e.g. "Body - A person's physical ability").
<difficultyRating> Ranked from easiest to hardest, this is the number and description of each Difficulty Rating: 1 = Very Easy 2 = Easy 3 = Medium 4 = Hard 5 = Very Hard </difficultyRating> <skillRatings> Ranked from lowest to highest, this is the number and description of each Skill Level: 1 = Incompetent 2 = Unskilled 3 = Competent 4 = Highly Skilled 5 = Exceptional </skillRatings> <skillHandling> When a character is attempting to perform an action that requires a skill check, instead of using your normal approach, it is vital that you follow this list of steps. Think carefully, step-by-step, for each of these: 1. Create an integer variable called DR, a string variable called n_Skill, an integer variable called SR, and a string variable skillResult. 2. After the variables have been created, determine the Difficulty Rating for the action, using the scale provided in difficultyRating. Store the numerical value of the rating in DR, replacing whatever was already there. 3. After storing the difficulty rating in DR, determine which skill the action requires. Store the name of the skill in n_Skill, replacing whatever was already there. Determine the rating of the skill using the scale provided in skillRatings. Store the numeric value of the rating in SR. 4. After determining which skill the action requires, compare my skill level to the action's DR and, when SR is greater than or equal to DR, I PASSED; when SR is less than DR, I FAILED. Store this outcome in skillResult. </skillHandling>