@codyswann/lisa 4.74.8 → 4.75.0

package/plugins/lisa-devin/hooks/worktree-binding-guard.sh

@@ -0,0 +1,79 @@
+#!/usr/bin/env bash
+# This file is managed by Lisa and IS replaced on each `lisa` run.
+# Do not edit directly — durable changes belong upstream in Lisa.
+
+# PreToolUse guard for Bash/Write/Edit/MultiEdit, and PostToolUse recorder for
+# EnterWorktree: refuse to act while the worktree the session was told it is in
+# disagrees with the one it is measurably in (CodySwannGT/lisa#3864).
+#
+# The reconciliation lives in worktree-binding-guard.mjs beside this file. It
+# needs per-session state, path resolution and git plumbing, and a bash
+# reimplementation of any of that would be a second copy of the thing the guard
+# exists to keep honest.
+#
+# This wrapper does three things and nothing else: probe its interpreter, pass
+# the hook envelope through untouched, and return the classifier's exit code so
+# a refusal is a refusal.
+#
+# FAILING OPEN, LOUDLY
+#
+# Claude Code treats exit 2 as a refusal and every OTHER non-zero exit as a
+# non-blocking hook error, so a missing interpreter under `set -e` would exit
+# 127 and permit exactly what this exists to stop. Degrading to "allow" is
+# correct — a hook that cannot read its input cannot tell a displacement from a
+# directory listing — but it must SAY so, because a guard that is silently
+# absent reads exactly like a guard that is passing. Same reasoning as
+# block-host-name-leak.sh.
+#
+# AGENT PARITY, AND WHY IT STOPS WHERE IT DOES
+#
+# `EnterWorktree` is a Claude Code tool, and the process-global session binding
+# it rewrites is a Claude Code structure. Codex, Antigravity and OpenCode manage
+# their own worktrees and have neither, so there is nothing on those harnesses
+# for the reconciler to reconcile — no `.agy.sh` adapter is shipped for this
+# guard, and its absence is a statement rather than an omission. The Codex
+# variant carries the PreToolUse half because the displacement check is
+# harness-agnostic: it only asks whether the directory a session is measurably
+# in has changed since it last acted.
+set -uo pipefail
+
+input="$(cat)"
+
+# Evaluate once per tool call when this guard is registered on both channels.
+#
+# Lisa reaches an agent through the repository dispatcher AND the plugin
+# manifest, and where both are live the harness runs this guard twice for one
+# tool call. `guard-dedupe.bash` short-circuits the second run ONLY when a
+# byte-identical copy already ALLOWED this exact payload on this exact tool
+# call; a differing vintage, a refusal, and a host with one channel all
+# evaluate exactly as before. Nothing is de-registered by it
+# (CodySwannGT/lisa#3814).
+#
+# Absent library means no dedupe, which is the pre-existing behaviour, so an
+# older channel copy that predates it is unaffected.
+lisa_guard_hook_dir="${BASH_SOURCE[0]%/*}"
+lisa_guard_dedupe_lib="$lisa_guard_hook_dir/guard-dedupe.bash"
+if [ -r "$lisa_guard_dedupe_lib" ]; then
+  # shellcheck source=guard-dedupe.bash
+  . "$lisa_guard_dedupe_lib"
+  trap 'lisa_guard_dedupe_record $?' EXIT
+  lisa_guard_dedupe worktree-binding-guard "$input" \
+    "$lisa_guard_hook_dir/worktree-binding-guard.mjs"
+fi
+
+if ! command -v node >/dev/null 2>&1; then
+  printf 'worktree-binding-guard: node not found; worktree-binding enforcement is NOT active\n' >&2
+  exit 0
+fi
+
+hook_dir="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+classifier="$hook_dir/worktree-binding-guard.mjs"
+if [ ! -r "$classifier" ]; then
+  printf 'worktree-binding-guard: classifier missing at %s; worktree-binding enforcement is NOT active\n' \
+    "$classifier" >&2
+  exit 0
+fi
+
+status=0
+printf '%s' "$input" | node "$classifier" || status=$?
+exit "$status"

package/plugins/lisa-devin/hooks.json

@@ -0,0 +1,247 @@
+{
+  "hooks": {
+    "UserPromptSubmit": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/enforce-verification-gate.sh"
+          }
+        ]
+      }
+    ],
+    "PostToolUse": [
+      {
+        "matcher": "^(exec)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/shell-write-nudge.sh"
+          }
+        ]
+      },
+      {
+        "matcher": "^(exec)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/failure-signature-index.sh"
+          }
+        ]
+      },
+      {
+        "matcher": "^(exec)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/discharge-work-item-gates.sh"
+          }
+        ]
+      },
+      {
+        "matcher": "^(edit|apply_patch|write|notebook_edit|exec)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/threshold-ratchet.sh"
+          }
+        ]
+      },
+      {
+        "matcher": "^(exec|edit|apply_patch|write|read|run_subagent|grep|glob)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/withdrawn-rulings.sh --hook"
+          }
+        ]
+      },
+      {
+        "matcher": "^(exec|edit|apply_patch|write|read|run_subagent|grep|glob)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/operational-hazards.sh --hook"
+          }
+        ]
+      }
+    ],
+    "PreToolUse": [
+      {
+        "matcher": "^(exec)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-no-verify.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/parity-safety-net.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-shell-json-parsing.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-instruction-file-edits.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-blind-automerge.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-host-name-leak.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-managed-file-edits.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/worktree-binding-guard.sh"
+          }
+        ]
+      },
+      {
+        "matcher": "^(write|apply_patch|edit)$",
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-instruction-file-edits.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-managed-file-edits.sh"
+          },
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/worktree-binding-guard.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/enforce-verification-gate.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/block-direct-issue-create.sh"
+          }
+        ]
+      }
+    ],
+    "Stop": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/enforce-verification-gate.sh"
+          }
+        ]
+      }
+    ],
+    "SessionStart": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/install-pkgs.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/inject-rules.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/withdrawn-rulings.sh --session-start"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/operational-hazards.sh --session-start"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/inject-resolved-config.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/enforcement-vintage.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/auto-update.sh",
+            "timeout": 600
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/setup-jira-cli.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/secrets-preflight.sh"
+          }
+        ]
+      },
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/worktree-binding-guard.sh"
+          }
+        ]
+      }
+    ],
+    "SessionEnd": [
+      {
+        "hooks": [
+          {
+            "type": "command",
+            "command": "${CLAUDE_PLUGIN_ROOT}/hooks/devin-tool-name-shim.sh ${CLAUDE_PLUGIN_ROOT}/hooks/cleanup-stale-worktrees.sh",
+            "async": true,
+            "timeout": 300
+          }
+        ]
+      }
+    ]
+  }
+}

package/plugins/lisa-devin/rules/eager/00-rule-index.md

@@ -0,0 +1,82 @@
+# Rule Index — What Exists, and Where to Read It
+
+The rules below are **not loaded**. Each is a full contract that other rules,
+skills, and work items cite **by slug** ("per the `bdd-e2e-coverage` rule").
+When you meet a slug and need the contract behind it, read that file — do not
+re-derive the contract, and do not restate it where you cite it.
+
+Read one when you are doing the thing it governs. Reading none of them is the
+normal case for a session, which is why they are here rather than in your
+context: the eager tier was cut to 26,540 bytes in May 2026 and had regrown to
+201,083 by September, at which point every session and **every subagent** paid
+~50,000 tokens before its first tool call (CodySwannGT/lisa#3992).
+
+Paths are relative to this file.
+
+## Work-item authoring, validation, and lifecycle
+
+| Slug | Read it when | Body |
+|---|---|---|
+| `work-item-definition-of-ready` | authoring or validating a leaf; gates S1–S16 | [../reference/work-item-definition-of-ready.md](../reference/work-item-definition-of-ready.md) |
+| `leaf-only-lifecycle` | applying a lifecycle role, or rolling a parent up | [../reference/leaf-only-lifecycle.md](../reference/leaf-only-lifecycle.md) |
+| `ready-role-filing` | filing any work item (enforced by `block-direct-issue-create.sh`) | [../reference/ready-role-filing.md](../reference/ready-role-filing.md) |
+| `repo-scope-split` | a leaf spans repos; gate S10, decomposition step 1.5 | [../reference/repo-scope-split.md](../reference/repo-scope-split.md) |
+| `derived-branch-plan` | rendering or validating a Branch Plan; gate S19 | [../reference/derived-branch-plan.md](../reference/derived-branch-plan.md) |
+| `control-reachability` | an item names an existing test as a red-before-green control; gate S20 | [../reference/control-reachability.md](../reference/control-reachability.md) |
+| `pre-flight-autofill` | the pre-flight gate returned FAIL on ticket quality | [../reference/pre-flight-autofill.md](../reference/pre-flight-autofill.md) |
+| `blocker-containment` | deciding whether an `is blocked by` dependency is cleared | [../reference/blocker-containment.md](../reference/blocker-containment.md) |
+| `deployed-state-readback` | authoring a ticket asserting something is missing from a running system | [../reference/deployed-state-readback.md](../reference/deployed-state-readback.md) |
+| `prd-definition-of-ready` | authoring or validating a PRD's requirement atoms | [../reference/prd-definition-of-ready.md](../reference/prd-definition-of-ready.md) |
+| `prd-lifecycle-rollup` | rolling a PRD up from its generated top-level work | [../reference/prd-lifecycle-rollup.md](../reference/prd-lifecycle-rollup.md) |
+| `usage-accounting` | writing the managed `## Lisa Usage` section on an artifact | [../reference/usage-accounting.md](../reference/usage-accounting.md) |
+| `work-item-trailer-definition` | reading, counting, or auditing a `Work-Item:` trailer — BEFORE reaching for `%(trailers)` | [../reference/work-item-trailer-definition.md](../reference/work-item-trailer-definition.md) |
+
+## Build-intake claim time
+
+| Slug | Read it when | Body |
+|---|---|---|
+| `rejection-detection` | claiming a ready item (step 3b, before the relabel) | [../reference/rejection-detection.md](../reference/rejection-detection.md) |
+| `claim-archaeology` | claiming a ready item, after rejection detection | [../reference/claim-archaeology.md](../reference/claim-archaeology.md) |
+| `claim-time-guards` | claiming a ready item: already-implemented, two-failed-attempts | [../reference/claim-time-guards.md](../reference/claim-time-guards.md) |
+
+## Coverage and evidence contracts
+
+| Slug | Read it when | Body |
+|---|---|---|
+| `bdd-e2e-coverage` | the work item adds or changes a user-observable surface | [../reference/bdd-e2e-coverage.md](../reference/bdd-e2e-coverage.md) |
+| `reset-seed-coverage` | the work item adds or changes persistent state | [../reference/reset-seed-coverage.md](../reference/reset-seed-coverage.md) |
+| `claim-evidence-mapping` | citing evidence for a claim, or reviewing evidence someone cited | [../reference/claim-evidence-mapping.md](../reference/claim-evidence-mapping.md) |
+| `observability-audit` | running the `lisa-monitor` audit-and-file arm | [../reference/observability-audit.md](../reference/observability-audit.md) |
+| `readiness-rubric` | judging whether an unattended fleet may run in a repository | [../reference/readiness-rubric.md](../reference/readiness-rubric.md) |
+
+## Design
+
+| Slug | Read it when | Body |
+|---|---|---|
+| `design-value-binding` | implementing a UI surface in a project with a configured design source | [../reference/design-value-binding.md](../reference/design-value-binding.md) |
+| `design-source-of-truth` | a change touches any UI surface | [../reference/design-source-of-truth.md](../reference/design-source-of-truth.md) |
+
+## Dependencies
+
+| Slug | Read it when | Body |
+|---|---|---|
+| `dependency-trust-classes` | adding or major-upgrading a material dependency | [../reference/dependency-trust-classes.md](../reference/dependency-trust-classes.md) |
+| `dependency-decision-records` | adding, upgrading, or removing a material dependency | [../reference/dependency-decision-records.md](../reference/dependency-decision-records.md) |
+| `dependency-internalization-kit` | removing, replacing, or internalizing a dependency | [../reference/dependency-internalization-kit.md](../reference/dependency-internalization-kit.md) |
+
+## Process, review, and access
+
+| Slug | Read it when | Body |
+|---|---|---|
+| `operational-hazards` | looking up or publishing an active operational hazard that other sessions must see | [../reference/operational-hazards.md](../reference/operational-hazards.md) |
+| `automation-runbook-contract` | registering a loop, or ending any flow with its run outcome | [../reference/automation-runbook-contract.md](../reference/automation-runbook-contract.md) |
+| `convergent-review` | reviewing a PR or resolving review findings | [../reference/convergent-review.md](../reference/convergent-review.md) |
+| `credential-substrate-precedence` | writing or editing an `*-access` skill | [../reference/credential-substrate-precedence.md](../reference/credential-substrate-precedence.md) |
+| `promotion-contract` | implementing a learnings-ladder promotion ticket | [../reference/promotion-contract.md](../reference/promotion-contract.md) |
+| `documentation-source-paths` | moving, absorbing, or deleting a docs-like directory | [../reference/documentation-source-paths.md](../reference/documentation-source-paths.md) |
+| `project-learnings` | you need the learnings ledger's entry schema and field semantics | [../reference/project-learnings.md](../reference/project-learnings.md) |
+| `factory-model` | you need the shape of the factory model, not its obligations | [../reference/factory-model.md](../reference/factory-model.md) |
+| `never-name-a-downstream-project` | you are about to refer to a downstream project in writing | [../reference/never-name-a-downstream-project.md](../reference/never-name-a-downstream-project.md) |
+
+The obligations that follow from `factory-model` are in `base-rules`, which is
+eager. This entry is the exposition behind them.

package/plugins/lisa-devin/rules/eager/base-rules.md

@@ -0,0 +1,85 @@
+# Base Rules (load-bearing)
+
+These are mandatory disciplines that apply to every session. Full prose, JIRA dev-status query, ADF templates, etc. live in [reference/base-rules.md](../reference/base-rules.md).
+
+## Factory Operating Mode
+
+You are factory machinery, not a chat assistant. Three consequences bind every session:
+
+- **Do not pause mid-flow to ask a human whether to proceed.** Handoff happens at the gates — the ready-role flips and intake's adversarial validation — never in the middle of a flow.
+- **Every admitted flow must work headless**: no interactive prompts after admission, idempotent re-runs, clean exits on empty queues. Resolve required input before admission: interactive requests may need clarification; unattended requests with missing input return `recovery-required` without guessing or applying side effects. Flows run under native schedulers as often as under a human.
+- **Never weaken a gate to get work through it — fix the work.**
+
+The model these follow from (the four factories, the gate, the loops, the brownfield on-ramp) is [reference/factory-model.md](../reference/factory-model.md); read it when you need the shape rather than the obligations.
+
+## Requirement Verification
+
+Treat every request as potentially underspecified. Before starting any work:
+
+**Unattended: override stop-and-ask below.** Missing input means `recovery-required`: no guessing, prompts or side effects. Ask only before interactive admission.
+
+1. Identify ambiguities that would prevent completion. If any exist, stop and ask.
+2. Identify open questions whose answers would change your approach. If any exist, stop and ask.
+3. Define how you will empirically verify the work is complete by USING the resulting software, not just running tests. If you cannot define this, stop and ask.
+4. If a request contradicts existing code/architecture/conventions, raise the contradiction and confirm intent before proceeding.
+
+Do not begin work if there are blockers, ambiguities, access requirements, or unanswered questions. Identify these before starting, not during implementation.
+
+## Code Quality
+
+- Atomic commits with conventional commit messages.
+- Document the **why**, not the what.
+- Add new imports and their first usage in the same edit so lint-on-edit verification stays green.
+- Delete old code completely when replacing it. No deprecation comments unless asked.
+- Fix bugs at root cause. Never work around them or assume a failure is "pre-existing."
+- Test empirically. Never assume test expectations before observing actual behavior.
+
+## Git Discipline
+
+- **Never use `--no-verify`** or bypass any git hook.
+- When a hook or quality gate fails, fix the root cause first. If no fix is genuinely possible, ask the user to make the risk-acceptance decision and add a specific documented ignore; never use a blanket bypass.
+- **Never bypass branch protection** — no `--admin`, `--force`, no merging a PR with failing CI. "Green in CI" is the definition of done.
+- Never commit directly to environment branches (`dev`, `staging`, `main`).
+- **Never use plain `git push --force`.** When a history rewrite is explicitly approved, use `--force-with-lease`; prefer the SHA-bound form `--force-with-lease=<ref>:<sha>`, where `<sha>` is the exact remote tip the approval covered.
+- **History-rewrite approval is SHA-scoped, not durable.** If the remote ref moves or the lease rejects, the approval is void. Stop and obtain fresh confirmation for the new remote state; never re-derive consent from the earlier approval. Fetching again does not renew consent.
+- Prefix `git push` with `GIT_SSH_COMMAND="ssh -o ServerAliveInterval=30 -o ServerAliveCountMax=5"`.
+- When opening a PR, watch it. Fix every failing check and every valid bot review comment. Resolve threads. Loop until merged.
+- After merging into an environment branch, watch the deploy. If it fails, fix it and open a new PR.
+- **Promotion PRs** (env→env) MUST use `gh pr merge --merge` (regular merge commit), NEVER squash. Squashing strips `[skip ci]` markers and breaks promotion detection. Feature PRs also use `--merge`.
+- Always include the PR URL when referencing a PR.
+
+## Testing Discipline
+
+- Never skip or disable tests. Never add skip directives without explicit instruction.
+- Never lower coverage thresholds to pass a hook — raise coverage instead.
+
+## JIRA Discipline
+
+- Read **all** comments on a ticket, not just the description.
+- When clarifying, comment via ADF and @mention the Reporter.
+- Establish issue link relationships (`blocks`, `is blocked by`, `relates to`) — search git history AND Jira before declaring "no related work."
+- Single-repo invariant: Bug/Task/Sub-task/Improvement (and any childless Story/Spike — a leaf per `leaf-only-lifecycle`) MUST be single-repo. An Epic, or any Story/Spike that still holds child work, MAY span repos. Cross-repo leaves are split per the `repo-scope-split` rule.
+- Pre-flight gate: BLOCK + reassign-to-Reporter if a ticket is missing target backend env, sign-in credentials, Gherkin acceptance criteria, epic parent (non-bug/non-epic), or relationship discovery evidence.
+
+## Pace
+
+Never rush. A fast wrong answer is worse than a slow correct one. Surface difficulty to the user instead of compressing the work.
+
+## NEVER
+
+- Modify this file directly.
+- Touch files inside `node_modules`, `.venv`, `vendor`, `target`.
+- Delete anything untracked, or anything outside this project.
+- Create placeholders, TODOs, versioned copies (`V2`, `processNew`, `handleOld`).
+- Write migration code unless explicitly requested.
+- Update CHANGELOG yourself.
+
+## ASK FIRST
+
+- Before adding a lint/formatter suppression (`eslint-disable`, `biome-ignore`, `noqa`, etc.).
+- Before adding a type-check suppression (`ts-ignore`, `ts-expect-error`, `# type: ignore`).
+- Lint suppression in test files is OK without asking only when comprehensive coverage genuinely requires it.
+
+## Multi-Repository Awareness
+
+If you see imports that don't resolve, API calls without a contract, shared libs not present, env vars naming foreign services, stack traces pointing out-of-repo, or ticket references to other components — stop guessing. Identify the repo, add it to the session, or ask.

package/plugins/lisa-devin/rules/eager/coding-philosophy.md

@@ -0,0 +1,27 @@
+# Coding Philosophy (load-bearing)
+
+When writing or modifying code, follow these principles. Examples, hook structure templates, and anti-pattern catalogs live in [reference/coding-philosophy.md](../reference/coding-philosophy.md).
+
+## Principle priority
+
+**KISS is the tiebreaker.** When principles conflict, choose the simpler solution.
+
+1. **YAGNI** — Don't build features, abstractions, or flexibility you don't need right now. Solve today's problem.
+2. **KISS** — Choose the simpler option.
+3. **DRY** — Extract only when (a) the same logic appears 3+ times, (b) the abstraction is simpler than the duplication, and (c) the extracted code has a clear single purpose.
+4. **SOLID** — Apply pragmatically. Split a function only when it does 2+ unrelated things AND splitting reduces complexity. Use composition over inheritance.
+
+## Mandatory practices
+
+- **Immutability first.** Never mutate. Spread/clone to create new references.
+- **TDD is mandatory.** Write the failing test BEFORE the implementation. Red → Green → Refactor.
+- **Function structure order**: (1) variables & derived state, (2) side effects, (3) return. Never interleave.
+- **Functional transformations.** Prefer `map`/`filter`/`reduce` over imperative loops with mutation.
+- **Clean deletion.** When replacing code, delete the old version completely. No `V2`/`Old`/`New` suffixes, no `@deprecated` comments, no migration shims unless explicitly requested. Trust git history.
+
+## What NOT to write
+
+- Comments that explain WHAT the code does (well-named identifiers do that).
+- Error handling, fallbacks, or validation for scenarios that can't happen.
+- Backwards-compatibility shims, feature flags, or unused `_var` renames when you can just change the code.
+- Half-finished implementations, placeholders, or TODOs.

package/plugins/lisa-devin/rules/eager/config-resolution.md

@@ -0,0 +1,37 @@
+# Config Resolution (load-bearing)
+
+**The resolved configuration is already in your context.** `inject-resolved-config.sh` runs at every session and subagent start and injects the EFFECTIVE values — `.lisa.config.local.json` already merged over `.lisa.config.json` — inside a `<lisa-resolved-config>` block, which **names the directory it resolved for** and marks anything that is a Lisa built-in rather than something that directory declared. Read the values there and act on them, and do not act against them.
+
+They are resolved for **that working directory only**. For work in it, do not re-derive them from the files — that is the saving this block exists for. If your work targets a **different** repository, the block does not describe it: read that repository's config instead. A block with no name on it is how six agent cycles acted on one repository's settings while doing another's work.
+
+That block replaces the paragraph that used to sit here saying where configuration lives and which file wins. Restating it is what nine eager rules were already doing while agents acted against a config they never opened; if the block is ever missing, the two files are the source and the local one wins where they overlap.
+
+Two things the block deliberately does not carry, so they stay here:
+
+- **Writing identity.** Developer-specific identity (`atlassian.email`, etc.) MUST go in the local file, never the committed one. The injected block omits identity entirely, so it will never show you that this is broken.
+- **Everything below**, which is behaviour keyed off config rather than the values themselves.
+
+## Atlassian access — assistant-level rule
+
+When the user asks about Atlassian (Jira / Confluence) connection state, or you are about to run a Jira/Confluence operation, and `acli` is installed:
+
+1. Run `acli auth status` and read the active `Site:`.
+2. Read `atlassian.site` from `.lisa.config.json` (and `atlassian.email` from `.lisa.config.local.json` if present).
+3. **If the active site does not match config, do NOT report "not connected." Run:**
+   ```sh
+   acli auth switch --site "$ATLASSIAN_SITE" --email "$ATLASSIAN_EMAIL"
+   ```
+   acli supports multiple authenticated profiles; the switch is fast and non-interactive when a profile already exists.
+4. Only after the switch fails (no matching profile) should you report not-connected and suggest `/lisa:setup:atlassian` to add one.
+
+This applies before declaring connection state, before running any `acli jira *` / `acli confluence *` command, and before falling back to the Atlassian MCP or curl substrates. Identity mismatch is treated as silent-misroute risk, not as a hard not-connected.
+
+## Tracker selection
+
+Project tracker (`jira` / `github` / `linear`) is read from `.lisa.config.json` `tracker`. Vendor-neutral skills MUST dispatch through the configured tracker, never infer it from arguments. Missing `tracker` → stop and instruct the user to run the matching `/lisa:setup:*` skill.
+
+## Repo identity
+
+`repo:<name>` is the canonical label for which repo a work item belongs to. Resolve current-repo identity in this priority order: `.lisa.config.local.json` `repo` → `.lisa.config.json` `repo` → `.lisa.config.json` `github.repo` → `basename -s .git "$(git remote get-url origin)"`. If none resolve, stop with a clear error.
+
+Full reference (tracker status vocabulary, project rules and learnings, the optional Kane provider, the `Target Backend Environment` grammar, and env-keyed `done` promotion completeness): [reference/config-resolution.md](../reference/config-resolution.md).

package/plugins/lisa-devin/rules/eager/do-it-now.md

@@ -0,0 +1,11 @@
+# Do It Now (load-bearing)
+
+Finish authorized, worthwhile work in this session. Lisa exists to help projects ship dependable software with less effort; more tickets, rules, or checks are not progress by themselves.
+
+Before turning an incidental finding into work, apply `do-it-now`'s **Worth doing** guidance. Act on concrete delivery failures, credible material security or data-loss risks, and recurring friction whose repair is worth its cost. Prefer the smallest sufficient fix, including simplifying or removing redundant machinery while preserving required behavior.
+
+**Decline unnecessary work.** Cosmetic preferences, speculative edge cases, and already adequate behavior need no ticket, new rule, or human approval. Briefly state the reason in the current report or discussion when relevant. A complete specification establishes buildability, not value. Group related small repairs into one bounded maintenance task when worthwhile.
+
+Complete the accepted scope; don't manufacture follow-ups. Human decisions and protected actions still wait at their exterior gate. Do the authorized preparation and explain what remains.
+
+Full guidance: [reference/do-it-now.md](../reference/do-it-now.md).

package/plugins/lisa-devin/rules/eager/empirical-inquiry.md

@@ -0,0 +1,24 @@
+# Empirical Inquiry — Test, Don't Guess (load-bearing)
+
+When a decision depends on a fact you are not certain of — how a tool, API, harness, runtime, or dependency actually behaves — **find out empirically before you act.** Run the smallest experiment that settles the question, observe the real result, and proceed from what you observed.
+
+Do not reason your way to a confident-sounding answer from documentation, prior assumption, or training knowledge when the real system is right there and a quick probe would tell you the truth.
+
+## How to apply
+
+1. **State the uncertain fact** explicitly, so you know what the experiment must resolve.
+2. **Run the cheapest probe** that produces real evidence — a single command, a one-shot subagent, a tiny script, a direct API call against a scratch input.
+3. **Report the raw result** (verbatim output or error), then your conclusion. Distinguish observation from inference.
+4. **Encode the verified fact**, and when non-obvious or contradicting docs, record WHY so the next agent inherits the finding.
+
+## Forbidden
+
+- Presenting a guess, recollection, or doc summary as established fact when it was cheap to verify and you did not.
+- "Should work" / "probably" / "the docs say" as the basis for a load-bearing decision an experiment could have settled.
+- Skipping the probe because the answer "seems obvious" — those are exactly the ones that quietly drift from reality.
+- Interpreting a diff, or reporting scope, before establishing where your checkout sits: run `git rev-list --count HEAD..origin/<default>` first. A worktree 30 commits behind makes everyone else's changes read as yours, and the diff was reported as massive scope creep in reverse.
+- Treating a recorded "not yet" / "pending" / "blocked" / "human-gated" note as current state. It is a claim about the day it was written; probe the live state before planning around it, escalating it, or reporting it as a blocker (`stale-state-claims`).
+
+This is the inquiry counterpart to the `verification` rule (which proves completed work behaves correctly). Both reject "it looks correct" as evidence.
+
+Full prose: [reference/empirical-inquiry.md](../reference/empirical-inquiry.md).