HAIL/HAIL_Schema_v1.0.yaml at main · human-ai-ledger/HAIL · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
# HAIL Session Ledger Schema v1.0 (YAML)
# Machine-readable structure for traceable Human–AI collaboration sessions.
# Author: Evan P. Troendle
# Date: 2026-02-19
#
# Captures:
#   - Sign-On Metadata (timestamps, model parameters, user ID, GDPR settings)
#   - Seed Block (identity, cognitive style, ethical/creative orientation, session purpose)
#   - Goal Statement (objectives and constraints)
#   - Checkpoints (iterative AI/human contributions, reflections, prompt histories, optional notes and decision changes)
#   - Final Reflection (aggregate session summary)
#   - Sign-Off Metadata (end timestamp, user confirmation)
# Defines the machine-readable structure for HAIL session ledgers, supporting auditability, provenance, and process traceability in human–AI collaboration.

metadata:
  schema_name: "HAIL Session Ledger"
  version: "1.0"
  description: "Machine-readable structure for traceable Human–AI collaboration sessions. Captures sign-on metadata, seed block, goal statement, iterative checkpoints, final reflection, and sign-off metadata. Supports auditability, process traceability, and automated validation."
  author: "Evan P. Troendle"
  date: "2026-02-19"
  citation: "Troendle, Evan P. (2026). Human–AI Ledger (HAIL): A Structured Workflow for Traceable, Reproducible Human–AI Collaboration. Zenodo. https://doi.org/10.5281/zenodo.17597705"
  version_history:
    - version: "1.0"
      date: "2026-02-19"
      changes: "Initial release of schema with normative session ledger structure."
Session_Ledger:
  Sign_On_Metadata:
    Sign_On_Timestamp: string  # ISO-8601 format, e.g. 2026-02-19T14:30:00Z
    Model_Version: string
    Parameters:
      # AI sampling / generation parameters; optional and model-agnostic
      temperature: float|null        # randomness control
      top_p: float|null              # nucleus sampling cutoff
      top_k: int|null                # max number of candidate tokens
      max_tokens: int|null           # maximum tokens generated
      stop_sequences: list|null      # optional sequences to halt generation
      presence_penalty: float|null   # discourages repeated tokens
      frequency_penalty: float|null  # discourages frequent token repetition
      seed: int|null                 # optional random seed
      # Stores any extra model-specific parameters not listed above (e.g., new sampling methods, fine-tuning metadata, embeddings)
      additional_params: map|null
    User_ID: string
    GDPR_Safe_Mode: boolean

  Seed_Block:
    Identity: string
    Core_Values: list
    Cognitive_Style: string
    Communication_Style: string
    Ethical_Orientation: string
    Creative_Orientation: string
    Purpose: string

  Goal_Statement:
    Goal: string
    Constraints: string

  Checkpoints:
    - Checkpoint_ID: int
      Checkpoint_Timestamp: string  # ISO-8601 format, e.g. 2026-02-19T14:31:00Z
      AI_Contribution: string
      Human_Contribution: string
      Reflections: string
      Notes: string|null
      User_ID: string|null  # optional; include only for multi-user sessions

  Full_Prompt_Summary: string
  Final_Reflection: string

  Sign_Off_Metadata:
    Timestamp: string  # ISO-8601 format, e.g. 2026-02-19T14:35:00Z
    User_Confirmation: boolean