@squadrules/mcp 5.12.1 → 5.13.1

package/dist/embed-docs/mem/README.md

@@ -19,7 +19,7 @@ observable verification over polished but weak wording.
 **Autonomy (AI-ownership).** The bundle is written for autonomous execution:
 adapters resolve intent from context, run critic-loop in a loop until it returns
 `PASS`, and gate irreversible remote mutations with objective pre/post contracts
-(shell/mcp) rather than human confirmation. They do **not** stop to ask a human to
+(shell/forward) rather than human confirmation. They do **not** stop to ask a human to
 approve or acknowledge. Human escalation is reserved for genuinely ambiguous intent
 or a hard blocker (a required tool/auth that cannot be established automatically).
 
@@ -28,9 +28,9 @@ observable verification available:
 
 1. `shell` — for file/state/command verification. **`interpreter` is REQUIRED**
    (allowed: `bash`, `sh`, `python3`, `python`, `node`).
-2. `mcp` — **forward-links only**. Must use `tool_name: "forward"` with
-   `arguments.uri` matching `squadrules://adapter/<slug>`. Bare tool-name-only
-   mcp contracts are not valid.
+2. `forward` — **links to another adapter**. Must set `forward.target` matching
+   `squadrules://adapter/<slug>`. Bare MCP tool calls are not valid forward
+   contracts — delegate them inside a `subagent` (rung 3).
 3. `subagent` — for delegated work that requires multi-step reasoning or tool
    orchestration by another agent.
 4. `user_input` — only for genuine human decisions.
@@ -61,7 +61,7 @@ author-adapter (Author) → forward(squadrules://adapter/critic-loop) → critic
 | File | Slug | Purpose |
 |------|------|---------|
 | [challenge-type-guide.md](challenge-type-guide.md) | `challenge-type-guide` | Challenge type selection, interpreter choice, and verification patterns |
-| [protocol-linking-guide.md](protocol-linking-guide.md) | `protocol-linking-guide` | Layers vs chains, `activate` vs `forward` + slug, chain-root collapsing, MCP forward-link validation |
+| [protocol-linking-guide.md](protocol-linking-guide.md) | `protocol-linking-guide` | Layers vs chains, `activate` vs `forward` + slug, chain-root collapsing, forward-link validation |
 | [critic-loop-guide.md](critic-loop-guide.md) | `critic-loop-guide` | Critic-loop integration rules, mutation gates, solution check vs critic decision table |
 
 ## Cross-Adapter Linking
@@ -94,17 +94,17 @@ call forward with squadrules://adapter/implement-terraform and no solution to st
 Slugs are globally unique (enforced by `train`), so resolution is
 deterministic — no scoring, no ambiguity.
 
-### MCP forward-link contracts
+### Forward-link contracts
 
-All `mcp` contracts must be forward-links with arguments:
+All `forward` contracts must name their target adapter:
 
-```json
-{"contract":{"type":"mcp","mcp":{"tool_name":"forward","arguments":{"uri":"squadrules://adapter/code-review-policy"}},"required":true}}
+Example ```json
+{"contract":{"type":"forward","forward":{"target":"squadrules://adapter/code-review-policy"},"required":true}}
 ```
 
-The server validates that `solution.mcp.arguments` is a superset of the
-contract's `mcp.arguments` (subset/deep matching). This proves the agent
-called `forward` with the correct slug, not just any `forward` call.
+The server validates that the submitted `evidence.target` equals the contract's
+`forward.target`. This proves the agent called `forward` with the correct slug,
+not just any `forward` call.
 
 ### When to use `activate` vs `forward` + slug
 

package/dist/embed-docs/mem/adapter-migration.md

@@ -1,5 +1,5 @@
 ---
-version: "5.12.1"
+version: "5.13.1"
 slug: adapter-migration
 title: Adapter Migration Workflow
 ---
@@ -78,7 +78,7 @@ Export every adapter from the target space using bulk export:
 If the space has no adapters, report this and stop.
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Export all adapters from the target space using the export tool with all_adapters: true. Record the total count and extract slug, URI, and content for each adapter."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Export all adapters from the target space using the export tool with all_adapters: true. Record the total count and extract slug, URI, and content for each adapter."},"required":true}}
 ```
 
 ## Analyze and Transform
@@ -97,8 +97,8 @@ For each exported adapter, perform structural analysis and apply migrations.
    (do NOT auto-split; report as manual review needed).
 5. **Shell interpreter** — every shell contract must include `"interpreter"`
    field with an allowed value (`bash`, `sh`, `python3`, `python`, `node`).
-6. **MCP forward-link** — every mcp contract must use `tool_name: "forward"`
-   with `arguments.uri` matching `squadrules://adapter/<slug>`.
+6. **Forward link** — every `forward` contract must set `forward.target`
+   matching `squadrules://adapter/<slug>`.
 
 ### Transformation rules
 
@@ -114,7 +114,7 @@ For each adapter, produce:
 - `transformed_content`: the full updated markdown (when `needs_changes`)
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Analyze each exported adapter against structural checks (heading order, contract fences, frontmatter, line count, shell interpreter, mcp forward-link). Apply conservative transformations and produce per-adapter status, changes, warnings, and transformed content."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Analyze each exported adapter against structural checks (heading order, contract fences, frontmatter, line count, shell interpreter, forward link). Apply conservative transformations and produce per-adapter status, changes, warnings, and transformed content."},"required":true}}
 ```
 
 ## Report Changes
@@ -152,7 +152,7 @@ For each `needs_changes` adapter, apply the structural changes:
 Do NOT change the adapter's space assignment during migration.
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Execute the migration: for each needs_changes adapter, call tune (content-level fixes) or train with force_update (structural changes). Record success/failure per adapter."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Execute the migration: for each needs_changes adapter, call tune (content-level fixes) or train with force_update (structural changes). Record success/failure per adapter."},"required":true}}
 ```
 
 ## Verify and Report
@@ -187,7 +187,7 @@ scheme. Every adapter has:
 - Correct heading order (Activation Patterns first, Reward Signal last).
 - Proper ` ```json ` contract fences.
 - Shell contracts with `interpreter` field.
-- MCP contracts as forward-links with `arguments.uri`.
+- Forward contracts with a `forward.target` adapter URI.
 - Valid frontmatter with `slug` and optional `version`.
 - Been verified post-migration for structural compliance.
 

package/dist/embed-docs/mem/author-adapter.md

@@ -1,6 +1,6 @@
 ---
 slug: author-adapter
-version: "5.12.1"
+version: "5.13.1"
 title: Author Adapter (Create / Review / Refactor)
 ---
 
@@ -87,7 +87,7 @@ inferred operation and target, then continue. Escalate to the user only when
 the request is genuinely ambiguous and no reasonable default exists.
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Resolve the authoring operation (create/review/update) and target adapter from user context. Record the decision."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Resolve the authoring operation (create/review/update) and target adapter from user context. Record the decision."},"required":true}}
 ```
 
 ## Gather Requirements
@@ -121,7 +121,7 @@ Export the target adapter, read the current markdown, identify the exact gap or
 error, and plan the minimal change.
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Gather requirements for the resolved operation: collect pattern/details for create, audit checklist for review, or identify the specific issue for update."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Gather requirements for the resolved operation: collect pattern/details for create, audit checklist for review, or identify the specific issue for update."},"required":true}}
 ```
 
 ## Draft Markdown
@@ -167,7 +167,7 @@ For `review`: write a structured gap report to `$REVIEW_FINDINGS_FILE`.
 For `update`: apply the identified change and show a diff-style summary.
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Draft the adapter markdown (create), findings report (review), or corrected markdown (update) following all shared drafting rules."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Draft the adapter markdown (create), findings report (review), or corrected markdown (update) following all shared drafting rules."},"required":true}}
 ```
 
 ## Critic-Loop Review (create and update only)

package/dist/embed-docs/mem/bulk-insert-adapters-via-cli.md

@@ -1,6 +1,6 @@
 ---
 slug: bulk-insert-adapters-via-cli
-version: "5.12.1"
+version: "5.13.1"
 title: Bulk Insert Adapters via SQUADRULES CLI
 ---
 

package/dist/embed-docs/mem/challenge-type-guide.md

@@ -1,5 +1,5 @@
 ---
-version: "5.12.1"
+version: "5.13.1"
 slug: challenge-type-guide
 title: Challenge Type Selection Guide
 ---
@@ -46,7 +46,7 @@ code — use `shell` or `subagent` instead.
 
 **Mode before challenge type:**
 - **clarify / approve** → `user_input`
-- **chain to another adapter** → `mcp` (forward-link only)
+- **chain to another adapter** → `forward` (link to another adapter)
 - **execute or verify local state** → `shell`
 - **delegate to another agent** → `subagent`
 - **pure reasoning with no observable artifact** → `comment`
@@ -55,7 +55,7 @@ code — use `shell` or `subagent` instead.
 
 ```
 Does the step chain to another adapter via forward?
- ├─ YES → "mcp" forward-link (tool_name: "forward", arguments.uri: squadrules://adapter/<slug>)
+ ├─ YES → "forward" (target: squadrules://adapter/<slug>)
  └─ NO → Does the step produce or verify a file, command result, or system state?
           ├─ YES → "shell" (with interpreter)
           └─ NO → Does the step require human approval or choice?
@@ -68,7 +68,7 @@ Does the step chain to another adapter via forward?
 ### Challenge JSON Formats
 
 **shell** (interpreter is REQUIRED)
-```json
+Example ```json
 {"contract":{"type":"shell","shell":{"cmd":"<command>","interpreter":"bash","timeout_seconds":30},"required":true}}
 ```
 
@@ -76,31 +76,41 @@ Allowed interpreters: `bash`, `sh`, `python3`, `python`, `node`. Build tools
 (`terraform`, `docker`, `npm`, etc.) are **not** interpreters — wrap them in a
 `bash` command instead.
 
-**mcp** (forward-link only — bare tool_name checks are not valid)
-```json
-{"contract":{"type":"mcp","mcp":{"tool_name":"forward","arguments":{"uri":"squadrules://adapter/<slug>"}},"required":true}}
+**forward** (links to another adapter — bare MCP tool calls are not valid)
+Example ```json
+{"contract":{"type":"forward","forward":{"target":"squadrules://adapter/<slug>"},"required":true}}
 ```
 
-The server validates that `solution.mcp.arguments` is a superset of the
-contract's `mcp.arguments` (subset/deep matching), proving the agent called
-`forward` with the correct target slug.
+The server validates that the submitted `evidence.target` equals the contract's
+`forward.target`, proving the agent called `forward` with the correct target
+slug. Bare MCP tool calls belong in a `subagent.mcp` list (rung 3), never here.
 
 **user_input**
-```json
+Example ```json
 {"contract":{"type":"user_input","user_input":{"prompt":"<specific question>"},"required":true}}
 ```
 
-**subagent**
-```json
-{"contract":{"type":"subagent","subagent":{"instruction":"<task description>"},"required":true}}
+**subagent** (rung 2 — delegated judgment)
+Example ```json
+{"contract":{"type":"subagent","subagent":{"context":"<task description>"},"required":true}}
+```
+
+Use `subagent` when a step delegates work to another agent rather than executing
+it directly. The `context` field provides task-specific guidance; the mandatory
+delegation instruction is appended automatically. Proof is `evidence.report` —
+the subagent's actual report, not your summary of it.
+
+**subagent.mcp** (rung 3 — delegated MCP tool work; optional)
+Example ```json
+{"contract":{"type":"subagent","subagent":{"context":"Export all adapters and report slugs and versions","mcp":[{"tool_name":"export","arguments":{"all_adapters":true}}]},"required":true}}
 ```
 
-Use `subagent` when a step delegates work to another agent (human or AI) rather
-than executing it directly. The `instruction` field provides task-specific
-guidance. Evidence of completion is collected via the anti-cheat evidence gate.
+`subagent.mcp` names the real MCP tool calls the **subagent** must make. The tool
+transcript stays in the subagent's context and never reaches yours; proof is
+still `evidence.report`. Omit the field for rung-2 judgment work.
 
 **comment** (last resort — max 1 per adapter, closing ACK only)
-```json
+Example ```json
 {"contract":{"type":"comment","comment":{"min_length":50},"required":true}}
 ```
 
@@ -123,14 +133,14 @@ Every solution follows the same envelope:
 {"type":"shell","outcome":"success","evidence":{"exit_code":0,"stdout":"...","stderr":"","duration_seconds":1.2}}
 ```
 
-**mcp solution**
+**forward solution**
 ```json
-{"type":"mcp","outcome":"success","evidence":{"tool_name":"forward","arguments":{"uri":"squadrules://adapter/critic-loop"},"response":{}}}
+{"type":"forward","outcome":"success","evidence":{"target":"squadrules://adapter/critic-loop","success":true}}
 ```
 
 **subagent solution**
 ```json
-{"type":"subagent","outcome":"success","evidence":{"agent_id":"agent-uuid","artifacts":["path/to/output"],"summary":"Task completed"}}
+{"type":"subagent","outcome":"success","evidence":{"report":"<the subagent's actual report>"}}
 ```
 
 **comment solution**
@@ -225,8 +235,8 @@ shell, then call `critic-loop` for adversarial review.
 **Chain-link via forward + slug** — layer N's contract requires `forward` with a
 specific adapter slug. Use when the adapter exceeds 350 lines or needs
 multi-path routing. Example:
-```json
-{"contract":{"type":"mcp","mcp":{"tool_name":"forward","arguments":{"uri":"squadrules://adapter/implement-plan"}},"required":true}}
+Example ```json
+{"contract":{"type":"forward","forward":{"target":"squadrules://adapter/implement-plan"},"required":true}}
 ```
 
 ## Artifacts for Shell Contracts
@@ -242,7 +252,7 @@ rather than inlining it in the contract JSON:
 
 This keeps adapter markdown clean and makes scripts independently versionable.
 
-```json
+Example ```json
 {"contract":{"type":"comment","comment":{"min_length":30},"required":true}}
 ```
 
@@ -255,6 +265,6 @@ The agent can now:
 2. match challenge type to the step's mode
 3. choose the right interpreter (bash, python3, node)
 4. apply working-directory and security rules
-5. use forward-link mcp contracts for chain links
+5. use forward contracts for chain links
 6. use subagent contracts for delegated work
 7. apply stronger verification patterns instead of weak comment checks

package/dist/embed-docs/mem/critic-loop-guide.md

@@ -1,5 +1,5 @@
 ---
-version: "5.12.1"
+version: "5.13.1"
 slug: critic-loop-guide
 title: Critic Loop Integration and Mutation Gate Guide
 ---
@@ -83,7 +83,7 @@ review with critic-loop instead of adding a hoop.
 | Read data from API | No | Read-only |
 | Compute / transform data | No | Local operation |
 
-```json
+Example ```json
 {"contract":{"type":"comment","comment":{"min_length":30},"required":true}}
 ```
 
@@ -99,7 +99,7 @@ Split proofs by **reversibility cost**, not by step count:
 - **The aggregate proof must be a strict conjunction** that re-verifies every
   sub-condition (`A && B && C && echo OK`):
 
-```json
+Example ```json
 {"contract":{"type":"shell","shell":{"cmd":"test -f \"$OUT_A\" && test -f \"$OUT_B\" && grep -q \"$EXPECTED\" \"$OUT_A\" && echo OK","interpreter":"bash","timeout_seconds":10},"required":true}}
 ```
 
@@ -114,20 +114,19 @@ The cheapest proof is the precondition that stops an irreversible mistake before
 it happens — for example, confirm you are on a throwaway worktree and branch (not
 `main`, not the human's active checkout) **before** making changes:
 
-```json
+Example ```json
 {"contract":{"type":"shell","shell":{"cmd":"test \"$(git rev-parse --abbrev-ref HEAD)\" != main && git rev-parse --git-common-dir | grep -q worktrees","interpreter":"bash","timeout_seconds":5},"required":true}}
 ```
 
 ## How to Chain Critic-Loop into an Adapter
 
 Use `forward` + slug to invoke `critic-loop` at phase boundaries. The standard
-pattern is a `mcp` contract that requires calling `forward` with the
-`critic-loop` slug:
+pattern is a `forward` contract naming the `critic-loop` slug:
 
 ### After a planning phase
 
-```json
-{"contract":{"type":"mcp","mcp":{"tool_name":"forward","arguments":{"uri":"squadrules://adapter/critic-loop"}},"required":true}}
+Example ```json
+{"contract":{"type":"forward","forward":{"target":"squadrules://adapter/critic-loop"},"required":true}}
 ```
 
 ### After an implementation phase
@@ -153,7 +152,7 @@ Record `work` in the handoff file pointed to by `$WORK_MANIFEST`, as flat keys
 `kind:`, `root:`, then `items:` with one item per line — so the precondition
 stays a machine-checkable file test instead of prose:
 
-```json
+Example ```json
 {"contract":{"type":"shell","shell":{"cmd":"test -f \"${WORK_MANIFEST:-}\" || test -f \"${INPUT_FILE:-}\"","interpreter":"bash","timeout_seconds":3},"required":true}}
 ```
 
@@ -193,7 +192,7 @@ Every adapter that chains critic-loop **must** include a shell contract that
 fails when the verdict is not PASS. This is the mechanical guard that prevents
 weaker agents from proceeding after FAIL:
 
-```json
+Example ```json
 {"contract":{"type":"shell","shell":{"cmd":"head -1 \"$VERDICT_FILE\" | grep -qi '^PASS$'","interpreter":"bash","timeout_seconds":5},"required":true}}
 ```
 
@@ -225,10 +224,10 @@ objective, machine-checkable contracts instead:
 
 ### Pre-mutation readiness (objective, no human prompt)
 
-Assert the preconditions with a shell or mcp contract so the action only runs from a
+Assert the preconditions with a shell or forward contract so the action only runs from a
 valid state — never with a `user_input` "are you sure?" prompt:
 
-```json
+Example ```json
 {"contract":{"type":"shell","shell":{"cmd":"test -n \"$TARGET_REPO\" && test -n \"$TARGET_BRANCH\"","interpreter":"bash","timeout_seconds":5},"required":true}}
 ```
 
@@ -238,8 +237,8 @@ Confirm the mutation actually took effect by reading the remote state back.
 Use a `subagent` contract when the verification requires multi-step reasoning
 or tool orchestration beyond a single shell command:
 
-```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Verify the remote mutation took effect: read back the created/updated resource and confirm its state matches the intended outcome."},"required":true}}
+Example ```json
+{"contract":{"type":"subagent","subagent":{"context":"Verify the remote mutation took effect: read back the created/updated resource and confirm its state matches the intended outcome."},"required":true}}
 ```
 
 The guarantee comes from critic-loop (truth of intent) plus these objective
@@ -278,7 +277,7 @@ When drafting a new adapter, apply this checklist:
 8. Verify that `train`/`tune` calls include `review_evidence` from the critic.
 9. Ensure the adapter's Reward Signal describes what the critic protects.
 
-```json
+Example ```json
 {"contract":{"type":"comment","comment":{"min_length":80},"required":true}}
 ```
 

package/dist/embed-docs/mem/critic-loop.md

@@ -1,5 +1,5 @@
 ---
-version: "5.12.1"
+version: "5.13.1"
 slug: critic-loop
 title: Critic Loop
 ---
@@ -158,7 +158,7 @@ never guessed at.
 **Output:** resolved work inventory for the audit
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Resolve the work submission into reviewable material: read the manifest, resolve each item by kind from work.root, and record the resolved inventory with per-item status."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Resolve the work submission into reviewable material: read the manifest, resolve each item by kind from work.root, and record the resolved inventory with per-item status."},"required":true}}
 ```
 
 ## Audit
@@ -219,7 +219,7 @@ Any `missing` item — or any `partial` high-risk item — forces FAIL with
 **Output:** structured audit findings for the verdict step
 
 ```json
-{"contract":{"type":"subagent","subagent":{"instruction":"Perform bounded adversarial audit: extract claims, investigate with scope discipline, gather evidence by priority hierarchy, produce claim verdicts, review phase boundary, and check requirement coverage."},"required":true}}
+{"contract":{"type":"subagent","subagent":{"context":"Perform bounded adversarial audit: extract claims, investigate with scope discipline, gather evidence by priority hierarchy, produce claim verdicts, review phase boundary, and check requirement coverage."},"required":true}}
 ```
 
 ## Verdict

package/dist/embed-docs/mem/protocol-linking-guide.md

@@ -1,5 +1,5 @@
 ---
-version: "5.12.1"
+version: "5.13.1"
 slug: protocol-linking-guide
 title: Adapter Linking Guide
 ---
@@ -125,21 +125,20 @@ call forward with squadrules://adapter/implement-terraform ...
 Slugs are globally unique (enforced by `train`), so resolution is
 deterministic — no scoring, no ambiguity.
 
-## MCP forward-link contracts for chain links
+## Forward-link contracts for chain links
 
-All `mcp` contracts MUST be forward-links with `tool_name: "forward"` and
-`arguments.uri` pointing to the target adapter slug:
+All `forward` contracts MUST set `forward.target` to the target adapter slug:
 
-```json
-{"contract":{"type":"mcp","mcp":{"tool_name":"forward","arguments":{"uri":"squadrules://adapter/code-review-policy"}},"required":true}}
+Example ```json
+{"contract":{"type":"forward","forward":{"target":"squadrules://adapter/code-review-policy"},"required":true}}
 ```
 
-The server validates that `solution.mcp.arguments` is a superset of the
-contract's `mcp.arguments` (subset/deep matching). This proves the agent
-called `forward` with the correct slug, not just any `forward` call.
+The server validates that the submitted `evidence.target` equals the contract's
+`forward.target`. This proves the agent called `forward` with the correct slug,
+not just any `forward` call.
 
-Bare `mcp` contracts that check only a tool name (without `arguments.uri`)
-are **not valid** — every mcp contract must specify the forward target.
+A `forward` contract without `forward.target` is **not valid** — every forward
+contract must specify its target adapter.
 
 ## Chain-root collapsing
 
@@ -181,11 +180,11 @@ and directing the agent to start from the chain root if those inputs are missing
 | Personal adapter (MCP database) | **Nothing** — must be self-contained | Only via `activate` (runtime discovery) |
 
 Personal adapters use `subagent` or `comment` contracts to describe what the
-agent should do next, not `mcp forward` contracts to other personal adapters.
+agent should do next, not `forward` contracts to other personal adapters.
 A personal adapter that chains to `squadrules://adapter/some-personal-slug`
 will fail at runtime when that slug does not exist in the caller's MCP database.
 
-```json
+Example ```json
 {"contract":{"type":"comment","comment":{"min_length":30},"required":true}}
 ```
 
@@ -197,7 +196,7 @@ The agent can now:
 1. decide when to use layers vs chains
 2. choose `activate` vs `forward` + slug correctly
 3. use slug-based linking for all adapters (system and user)
-4. write valid mcp forward-link contracts with arguments
+4. write valid forward contracts with a target adapter URI
 5. implement `chain_root` collapsing with defense-in-depth safeguards
 6. document chaining back (`chain_root` + prerequisites) and chaining forward
    (`forward` with known slugs) for root + extension families

package/dist/embed-docs/mem/refine-search.md

@@ -1,5 +1,5 @@
 ---
-version: "5.12.1"
+version: "5.13.1"
 slug: refine-search
 title: Get help refining your search
 ---