# SREL/AI-1 Primitive Reference Table

The fifty-four named risk primitives of the SREL/AI-1 profile populate the matrix of nine dyadic groups by six compositional periods.  Each primitive is identified by its coordinate `<group>.<period>`, its three-to-five-character coding token, and its descriptive name.  The assignments below are illustrative; deployment-specific overrides are permitted subject to two constraints: (a) primitive assignments are mutually exclusive within their (group, period) coordinate, and (b) coding tokens are unique within a single SREL document.

## Groups I–IV (Model-anchored dyads)

| Coord  | Code | Name                     | Group / Period            |
| ------ | ---- | ------------------------ | ------------------------- |
| I.P1   | HAL  | Hallucination            | I (L–S) / P1 Primordial    |
| I.P2   | BIA  | Bias                     | I (L–S) / P2 Relational    |
| I.P3   | OVF  | Overfitting              | I (L–S) / P3 Synthetic     |
| I.P4   | UND  | Underfitting             | I (L–S) / P4 Mediated      |
| I.P5   | MD   | Model drift              | I (L–S) / P5 Social        |
| I.P6   | DRF  | Data drift               | I (L–S) / P6 Absolute      |
| II.P1  | LAT  | Latency risk             | II (L–P) / P1 Primordial    |
| II.P2  | ROB  | Reliability risk         | II (L–P) / P2 Relational    |
| II.P3  | DEP  | Deployment risk          | II (L–P) / P3 Synthetic     |
| II.P4  | INT  | Integration risk         | II (L–P) / P4 Mediated      |
| II.P5  | SCL  | Scaling risk             | II (L–P) / P5 Social        |
| II.P6  | CST  | Cost overrun             | II (L–P) / P6 Absolute      |
| III.P1 | ADV  | Adversarial attack       | III (L–A) / P1 Primordial   |
| III.P2 | JBR  | Jailbreak                | III (L–A) / P2 Relational   |
| III.P3 | PJI  | Prompt injection         | III (L–A) / P3 Synthetic    |
| III.P4 | API  | API abuse                | III (L–A) / P4 Mediated     |
| III.P5 | TOK  | Token theft              | III (L–A) / P5 Social       |
| III.P6 | EXT  | Data exfiltration        | III (L–A) / P6 Absolute     |
| IV.P1  | POL  | Policy failure           | IV (L–Ω) / P1 Primordial    |
| IV.P2  | GOV  | Governance failure       | IV (L–Ω) / P2 Relational    |
| IV.P3  | REG  | Regulatory risk          | IV (L–Ω) / P3 Synthetic     |
| IV.P4  | CMP  | Compliance risk          | IV (L–Ω) / P4 Mediated      |
| IV.P5  | OWN  | Ownership gap            | IV (L–Ω) / P5 Social        |
| IV.P6  | ETH  | Ethical risk             | IV (L–Ω) / P6 Absolute      |

## Groups V–VII (Data-anchored dyads)

| Coord   | Code | Name                   | Group / Period            |
| ------- | ---- | ---------------------- | ------------------------- |
| V.P1    | PII  | Personal data exposure | V (S–P) / P1 Primordial    |
| V.P2    | CON  | Consent risk           | V (S–P) / P2 Relational    |
| V.P3    | DAT  | Data leakage           | V (S–P) / P3 Synthetic     |
| V.P4    | DUP  | Duplicate data         | V (S–P) / P4 Mediated      |
| V.P5    | SRC  | Source risk            | V (S–P) / P5 Social        |
| V.P6    | DQL  | Data quality loss      | V (S–P) / P6 Absolute      |
| VI.P1   | MSL  | Mislabeling            | VI (S–A) / P1 Primordial    |
| VI.P2   | SYN  | Synthetic-data risk    | VI (S–A) / P2 Relational    |
| VI.P3   | EXP  | Explainability gap     | VI (S–A) / P3 Synthetic     |
| VI.P4   | ACC  | Access risk            | VI (S–A) / P4 Mediated      |
| VI.P5   | SUP  | Supply-chain risk      | VI (S–A) / P5 Social        |
| VI.P6   | UNC  | Uncertainty            | VI (S–A) / P6 Absolute      |
| VII.P1  | MON  | Monitoring gap         | VII (S–Ω) / P1 Primordial   |
| VII.P2  | LOG  | Logging gap            | VII (S–Ω) / P2 Relational   |
| VII.P3  | DOC  | Documentation gap      | VII (S–Ω) / P3 Synthetic    |
| VII.P4  | AUD  | Audit gap              | VII (S–Ω) / P4 Mediated     |
| VII.P5  | EVAL | Evaluation failure     | VII (S–Ω) / P5 Social       |
| VII.P6  | FBK  | Feedback gap           | VII (S–Ω) / P6 Absolute     |

## Groups VIII–IX (Harm-emergent and Harm-purposive dyads)

The following assignments for Groups VIII and IX are exemplary defaults subject to deployment-specific override.

| Coord   | Code  | Name                     | Group / Period            |
| ------- | ----- | ------------------------ | ------------------------- |
| VIII.P1 | REINF | Harm reinforcement       | VIII (P–A) / P1 Primordial |
| VIII.P2 | AMPL  | Deployment amplification | VIII (P–A) / P2 Relational |
| VIII.P3 | CASC  | Cascading harm           | VIII (P–A) / P3 Synthetic  |
| VIII.P4 | PROL  | Harm proliferation       | VIII (P–A) / P4 Mediated   |
| VIII.P5 | NORM  | Harm normalization       | VIII (P–A) / P5 Social     |
| VIII.P6 | IRR   | Harm irreversibility     | VIII (P–A) / P6 Absolute   |
| IX.P1   | MIS   | Mission drift            | IX (P–Ω) / P1 Primordial    |
| IX.P2   | VAL   | Value drift              | IX (P–Ω) / P2 Relational    |
| IX.P3   | GAM   | Goal gaming              | IX (P–Ω) / P3 Synthetic     |
| IX.P4   | SUB   | Purpose subversion       | IX (P–Ω) / P4 Mediated      |
| IX.P5   | COR   | Purpose corruption       | IX (P–Ω) / P5 Social        |
| IX.P6   | TRG   | Goal mistargeting        | IX (P–Ω) / P6 Absolute      |

## Excluded edge

The A–Ω edge (Emergence–Purpose) is excluded from per-edge scoring as the irreducible alignment edge.  This exclusion is structural: no tenth dyadic group is enumerated, and the alignment problem is treated as a structural property of the substrate rather than as a peer-level primitive entry within the matrix.

## Compound primitives

Compound primitives are reported under `[[compound]]` sections with constituents drawn from the 54 base primitives.  Illustrative compounds include ERR (error compounds), STR (stress compounds), VAL (validation compounds), RED (redundancy compounds), ROL (role compounds), TLS (tool compounds), DEL (delivery compounds), GOAL (goal compounds), REV (revision compounds), and MISU (misuse compounds).