manual_slop/docs/ideation/2026-06-12-intent-based-scripting-languages.md

Intent-Based Scripting Languages

Track: intent_dsl_survey_20260612 (initialized 2026-06-12) Date: 2026-06-12 Author: Tier 1 Orchestrator (sections 1, 3, 4, 5, 6, 7, Appendix); Tier 2 sub-agents (section 2 clusters 0, 1, 2, 3, 4) Status: Outline draft (phase 1 of 4)

What this is. A survey of intent-based scripting languages as a design philosophy, plus a proposed vocabulary (~40 verbs across 4 tiers) for a Meta-Tooling-facing intent DSL. The report is the foundation document for the user's nagent v2.2 (its "Future-Track Candidate #4" section) and for the future interpreter prototype (follow-up B track).

What this is NOT. Not an interpreter, not a bridge script, not Application-side function-calling, not XML/JSON record formats. The DSL is Meta-Tooling-side per docs/guide_meta_boundary.md — the format external agents (Gemini CLI, OpenCode) emit when invoking mcp_client.py tools.

---

1. The "Intent-Based" Design Philosophy

[STUB: 4 anchor claims — O'Donnell immediate-mode as the philosophical anchor; Onat/Lottes hardware-pipeline model as the truth the verbs must map to; CoSy open-vocabulary culture as the user-surface principle; Jofito intent-mapping as the framing that names the design philosophy. ~2-3 pages.]

2. Prior Art Survey (8 Clusters)

This section surveys the design lineage across 8 clusters. Each entry: 2-3 sentences on the design idea, 2-3 sentences on what we take from it (or, in cluster 3, what we explicitly reject). Every entry cites a specific source.

Cluster 0 — Immediate-Mode Paradigm (philosophical anchor)

[STUB: John O'Donnell's IMGUI/MVC essays. ~0.5-1 page.]

Cluster 1 — Concatenative (Forth family)

[STUB: Forth, ColorForth, KYRA/Onat, x68/Lottes, Joy, CoSy. ~1-1.5 pages.]

Cluster 2 — Array

[STUB: APL, K, BQN, Uiua. ~0.5 page.]

Cluster 3 — Intent-Mapping

[STUB: Jofito, jq, nagent's tag protocol (REJECTED as a model — we take the structured-protocol idea but not the XML angle brackets), Wasm. ~0.5-1 page.]

Cluster 4 — Meta-Tooling DSLs and Agent-Facing Languages

[STUB: mcp_dsl_20260606 placeholder, nagent's Bridge DSL, OpenAI function-calling, Anthropic tool-use. ~0.5 page.]

Cluster 5 — SSDL Shape Primitives

[STUB: 6 primitives + 7 modifiers per docs/reports/computational_shapes_ssdl_digest_20260608.md §1. The meta-vocabulary used to annotate the verbs in section 4. ~0.25 page.]

Cluster 6 — Project's Own Command DSL Precedents

[STUB: the 33 Command Palette commands per docs/guide_command_palette.md and src/commands.py. The DSL is a richer superset; "Everything" mode is a near-term use case. ~0.25 page.]

Cluster 7 — Data-Oriented Error Handling Convention

[STUB: Result[T] + ErrorInfo per conductor/tracks/data_oriented_error_handling_20260606/spec.md §3.3. The DSL's try/recover/sandbox/didyoumean verbs return Result[T]. ~0.25 page.]

3. The Grammar

[STUB: 14 primitives formalized from the user's math pseudocode (determinate/minor/matrix-transpose snippets). Each primitive: symbol, name, signature, one-line semantics, example, "borrowed from" note. Plus 3 known ambiguity flags (proc placement, ++ → += 1, m[row][column] → m[row, col]). Plus precedence rules (left-to-right for -> chains, () for grouping) and AI-fuzzing tolerance rules (CoSy-style modulo indexing, structured recovery anchors via {}, line/offset independence). Plus the error envelope (try { ... } recover { ... } returns Result[T]). ~2-3 pages.]

4. The 4-Tier Vocab (~40 Verbs)

Each verb: symbol, name, signature, one-line semantics, one example, "borrowed from" note, SSDL shape tag. Tier 2 and Tier 3 verbs also have a "maps to mcp_client tool" column.

Tier 1 — Math (~10 verbs)

[STUB: from the user's pseudocode. :=, stack { }, for x .. n, +, -, *, /, ^, sum, product, a[i,j], if/then.]

Tier 2 — Data-Oriented Pipeline (~12 verbs)

[STUB: Onat/Lottes/Jofito lineage. scan, select, filter, map, fold, sort, group, dedupe, arena { }, scatter, gather, pipe. The verbs that wrap the existing 45+ MCP tools.]

Tier 3 — Shell (~10 verbs)

[STUB: the OS surface. exec, open, read, write, close, path, env, wait, poll, cwd.]

Tier 4 — AI-Fuzzing Tolerance (~8 verbs — the novel contribution)

[STUB: the verbs that make the DSL work for AI agents that may fuzz verb names, indent inconsistently, or offset line references. fuzzy, try { ... } recover { ... }, sandbox { ... }, audit, didyoumean, span, offset, assumewide. The sandbox verb is O'Donnell's IEventTarget pattern applied to the DSL; the audit verb is the IEventTarget itself.]

5. Hardware Mapping (4 Anchor Claims)

[STUB: 4 anchor claims tying the vocab to actual hardware/software stages.

1. Onat/Lottes, hardware: the 2-register stack + magenta pipe + basic blocks + lambdas + preemptive scatter (per C:\projects\forth\bootslop\references\kyra_in-depth.md, forth_day_2020_in-depth.md, neokineogfx_in-depth.md, X.com - Onat & Lottes Interaction 1.png.ocr.md) → our ->, [ ], arena { }, scatter/gather.
2. O'Donnell, paradigm: the DSL's pipeline is immediate-mode in pipeline composition. Per https://johno.se/book/imgui.html. The -> chain has no "pipeline object" you can query, name, or pass around.
3. Forth/CoSy, syntax: concatenative syntax is immediate-mode in tokenization (whitespace-delimited, no precedence), evaluation (each verb pops args, pushes results), and parsing (no AST object retained after parse).
4. APL/K, data: array languages are immediate-mode in data representation. The DSL's for x .. n + result[row, col] inherits the "no array object" property.

~1-2 pages.]

6. AI-Agent Properties (10 Claims)

[STUB: 10 claims tying the DSL to the existing project's architecture so future tracks can build on it without re-deriving the design.

1. Domain = Meta-Tooling (per docs/guide_meta_boundary.md).
2. Runtime path = external agent → DSL text → bridge script → MCP → optional Hook API approval (per docs/guide_meta_boundary.md §"The Inter-Domain Bridges").
3. 3-layer security (per docs/guide_tools.md §"The MCP Bridge"): the parser rejects DSL statements that target tools outside the allowlist.
4. 4 memory dimensions (per conductor/tracks/nagent_review_20260608/nagent_review_v2_1_20260612.md §2.1): the DSL does not replace any memory dimension.
5. Stable-to-volatile cache ordering (per nagent v2.1 §2.2): the DSL's arena { } blocks are cache-friendly.
6. Result[T] envelope (per conductor/tracks/data_oriented_error_handling_20260606/spec.md): try/recover verbs return Result[T].
7. Command Palette 33 commands (per docs/guide_command_palette.md): the DSL is a richer superset; "Everything" mode is a near-term use case.
8. Hook API state fields (per docs/guide_state_lifecycle.md §"Hook API Surface"): DSL verbs route through _predefined_callbacks and _gettable_fields.
9. O'Donnell's IEventTarget pattern as the sandbox verb (per https://johno.se/book/mvc.html §"Writing to Model state").
10. O'Donnell's "reads are free" claim (per https://johno.se/book/mvc.html §"Reading Model state"): Tier 2 verbs are read-only and re-evaluable.

~2-3 pages.]

7. Open Questions for Follow-up B (≥6)

[STUB: open questions that the follow-up B track (interpreter prototype) must answer, plus the connection block to the intent_dsl_for_meta_tooling_20260608_PLACEHOLDER.

1. How does arena { } map to Onat's preemptive scatter? Block-as-tape vs wrapper-allocates-tape?
2. Where does "intent resolution" live? Per-verb option, per-block modifier, or global parser mode?
3. How does audit interact with Manual Slop's existing comms.log? Separate JSON-L or merged?
4. Does sandbox produce Result[T, ErrorInfo] (Fleury pattern)?
5. didyoumean recovery: parser feature or user-facing verb?
6. How does for x .. n interact with Tier 2's filter/map? Sugar or distinct?
7. How does sandbox map to Manual Slop's pre_tool_callback flow? New audit log or fold into existing?
8. Connection to intent_dsl_for_meta_tooling_20260608_PLACEHOLDER: minimum vocab subset for one round-trip end-to-end?

~1-2 pages.]

---

Appendix: Bibliography

[STUB: full file:line / URL references for all 8 prior-art clusters + the project's own references. Grouped by cluster. Each entry: 1 line with the source identifier and the file:line or URL.]