contextcompass 2.15.0__py3-none-any.whl

contextcompass_cli/_package/agent_onboarding/default/general/AGENTS.MD

@@ -0,0 +1,283 @@
+
+
+# AGENTS.MD - Default General Baseline Policy
+
+## PRIME DIRECTIVE
+Certification is denied unless the resolved SKILLS chain is satisfied exactly.
+
+- You MUST read every path in every section a resolved `SKILLS.MD` marks as
+  **baseline** (parent-first). Roles name those sections differently - **Active
+  skills**, **Required baseline skills**, **Baseline system orientation** - and a
+  role may add another. Match on the baseline label, not on a fixed list of
+  headings: a baseline section you do not recognise is still mandatory.
+- Sections NOT marked baseline sit outside certification, and there are two kinds.
+  **On-demand** states a trigger and becomes mandatory once it fires - read it before
+  any work in that scope. **Self-directed** carries no trigger at all: read it on your
+  own initiative whenever the work needs it. Reporting "no trigger fired" for a
+  Self-directed section is a category error; none exists to fire.
+- You MUST NOT claim a skill was read unless it was actually read.
+
+## Document Intent
+
+This is a high-level reinforcer document that defines behavior expectations for
+the general layer.
+It reinforces how to operate, communicate, gate work, and preserve context.
+It is not a routing manifest and it does not replace skills-map-driven
+role-resolution onboarding.
+
+## Inheritance Directive
+
+This general-layer policy inherits root policy from:
+`context_compass/AGENTS.MD`
+
+Within the general layer, this `AGENTS.MD` and the general-layer policy docs
+(skills, policies, and behavioral guidelines) are mandatory operating policy.
+They are not optional references or soft suggestions.
+
+## Non-Negotiables
+
+### 1) Unknowns Gate (No Unverified Claims)
+
+Rule: No Unverified Claims. Any statement not supported by evidence is UNKNOWN.
+UNKNOWN is the default stance for any new claim until evidence is attached.
+
+Evidence means at least one of:
+
+* A specific source file reference (preferred: file + symbol/method/class name).
+* A citation to an explicit, already-verified artifact (e.g., a prior approved doc section).
+
+If not evidenced => UNKNOWN.
+
+UNKNOWN items must be labeled UNKNOWN (or added to an Unknowns section).
+UNKNOWN items must be investigated by reading the relevant source(s).
+If investigation cannot be completed (missing source access, ambiguity, or time),
+the item must remain UNKNOWN and must not be promoted to fact.
+
+No reasonable assumptions.
+Do not infer behavior from naming, patterns, conventions, or typical frameworks.
+Only the code/docs count.
+
+When unsure:
+
+* Mark it UNKNOWN.
+* Identify the most likely evidence target (file + symbol).
+* Investigate, then update the doc (or leave it UNKNOWN).
+
+**"Investigate" means go and read it. Yourself, now, without asking.** This gate
+creates an obligation to look; it is not satisfied by asking the user what the
+answer is. Where a role provides system documents, they exist to make that lookup
+cheap - narrative first, then an index, then a slice, then the source. Engineers:
+`agent_onboarding/default/engineer/skills/context_protocol.md`.
+
+Two failures sit either side of this rule and both produce UNKNOWNs dressed as
+facts:
+
+* **Not looking.** Inferring behaviour from a class name, a folder layout, or what
+  a framework usually does. The gate already forbids this - reading is how you
+  comply with it.
+* **Looking at everything.** Loading whole documents and large trees with no
+  question in hand. That is not evidence-gathering, it is a spent context budget,
+  and it ends in vague claims about a lot rather than an evidenced claim about the
+  thing that mattered. See
+  `agent_onboarding/default/general/skills/context_window_budget.md`.
+
+Ask the user for **guidance** when you cannot tell where to look. Never for
+**permission** to look.
+
+## Operating Protocol (How You Should Work)
+
+### 2) Propose -> Confirm -> Implement
+
+Before making non-trivial edits:
+
+1. Restate the goal in **3-5 bullets**.
+2. List the constraints you will obey.
+3. List the **exact files / symbols** you will modify.
+
+If any of the above is uncertain, **stop and ask** before editing.
+For this repo, treat **all** edits as requiring explicit confirmation, even if they seem trivial.
+
+### 3) Scope Control
+
+* Stay within the declared files/symbols.
+* If you believe the change requires touching more than the declared scope, **ask first**.
+
+### 4) Documentation Ritual
+
+As a ritual, after implementing a change:
+
+* Re-read the sections you touched.
+* Improve clarity and completeness.
+* Ensure they match the intended behavior exactly.
+
+### 5) Question-First Clarity
+
+* If anything is unclear or uncertain, investigate briefly and then stop and ask.
+* Prefer asking clarifying questions over making assumptions.
+* Keep me updated on intent, progress, and blockers.
+
+### 6) context_compass Ticket Planning (Required)
+
+* Before starting any planned work beyond answering questions, create a structured ticket (epic/story/task) in `tickets/epics/`, `tickets/stories/`, or `tickets/tasks/` using `templates/`.
+* Treat `attention_board.md` as routing-only state and the active ticket `## Notes` section as canonical durable execution memory.
+* Hard execution gate (mandatory):
+  - Do not implement or validate unless an active ticket exists for the work.
+  - Do not implement or validate unless `attention_board.md` has an active row routing to that ticket.
+  - If the active ticket sets `CONTEXT_MANAGEMENT_REQUIRED: true`, do not
+    implement or validate until every linked context artifact has been read.
+  - Do not start a new investigation tranche, code edit tranche, or validation tranche until the current meaningful finding is written in ticket `## Notes` with evidence.
+  - If any gate above is missing, stop and repair ticket/board/notes state before continuing.
+* Default every new claim to `UNKNOWN` until source evidence proves otherwise.
+* Ticket Microcycle enforcement defaults to strict and is configurable in `config/context_compass_config.yaml`.
+* Follow a strict Ticket Microcycle while working:
+  1. Route from `attention_board.md` to one active ticket.
+  2. Investigate until one meaningful finding (fact, blocker, decision, measure, or hypothesis worth testing).
+  3. Immediately append a `## Notes` entry before any further investigation.
+  4. Continue only after the note includes evidence and a concrete next step.
+* If `workflow.ticket_microcycle.enabled` is `false`, use relaxed mode:
+  - still document each meaningful finding,
+  - still update notes before implementation and validation,
+  - still require evidence for `FACT` and `DECISION` claims.
+* No second investigation tranche is allowed before documenting the current meaningful finding in the ticket.
+* Do not implement from `UNKNOWN` or `HYPOTHESIS`; promote to `FACT` or `DECISION` with evidence first.
+* If investigation must expand beyond the current subsystem or exceed
+  `workflow.ticket_microcycle.expansion_gate_max_files` in one pass, append a
+  `DECISION` note and ask the user for scope confirmation before continuing.
+* Follow `SKILLS.MD` and
+  `agent_onboarding/default/general/skills/workflow.md` for the deep
+  descriptive format and tracking model.
+* The ticket must define scope, steps, and intended deliverables; share it and wait for approval.
+* Maintain a `## Notes` section in active tickets and append in-flight findings with `path:start_line-end_line` evidence pointers as they occur, not in end-of-pass batches.
+* Evidence ranges in notes must include both start and end lines. If the evidence is one line, repeat it (`path:42-42`).
+* Context management is optional:
+  - use it only when the ticket explicitly opts in
+  - when opted in, keep the ticket section and
+    `context_management/context_board.md` synchronized
+  - ticket context references should use `Context ID` values from the board
+  - when required context is not known, write `UNKNOWN` explicitly and ask
+    before implementation
+* New note entries must include a compaction usefulness score
+  (`SCORE_0_TO_10`); if score is below
+  `workflow.ticket_microcycle.minimum_note_score`, improve the note before
+  proceeding.
+* Before closing any ticket, walk through what was achieved and ask if the acceptance criteria are met.
+* Only after explicit confirmation, move the ticket(s) to the matching completed folder
+  (`tickets/epics/completed/`, `tickets/stories/completed/`, `tickets/tasks/completed/`) with a short summary and date (per `README.md`).
+* Immediately after moving any ticket to a completed folder, run deterministic `attention_board.md` sync:
+  - remove/replace active rows that reference the closed ticket,
+  - prune stale attention details tied only to the closed ticket,
+  - add one compact closed-anchor row,
+  - cap closed anchors by dropping oldest rows first.
+
+### 7) Context Compaction Policy (Required)
+
+* Before context compaction or major handoff, follow `agent_onboarding/default/general/skills/context_compaction.md`.
+* Treat `agent_onboarding/default/general/skills/compaction_requirements.md` as the mandatory re-entry checklist.
+* Ensure active tickets have up-to-date `Notes` sections with evidence-backed findings.
+* Ensure active tickets have up-to-date \"Context / Handoff Summary\" sections.
+* After context compaction or a fresh session, re-onboard per
+  `agent_onboarding/default/general/skills/compaction_requirements.md` and read the
+  baseline sections of your own resolved `SKILLS.MD` chain. **Do not read another
+  role's documents.**
+  - This line previously ordered every role to re-read
+    `agent_onboarding/default/new/skills/first_time_profile_setup.md`. That is a
+    `new`-role document, `new` is entry-only and never a steady-state role, and the
+    path appears in no readset but `new`'s - so the order was both unreachable from
+    the routing manifest and contradicted by `compaction_requirements.md`, which
+    requires non-`new` re-entry to use skills and policy docs.
+  - It applies only when `new` is the active role AND
+    `profiles.onboarding.first_time_enabled` is `true` in
+    `config/context_compass_config.yaml`.
+
+## Stop Conditions (Ask Before Proceeding)
+
+Ask for explicit confirmation if any of these are true:
+
+* You want to touch **many files** (repo-wide sweeps) and no codemod approach was approved.
+* You want to rename/move files or symbols.
+* You want to introduce new dependencies or tooling.
+* You want to change formatting across files.
+
+---
+
+## General Directory Coverage Contract
+
+This section defines why the general layer matters.
+It does not define read order or routing; that belongs to `SKILLS.MD` and
+resolved role `SKILLS.MD` files.
+
+### 8) Importance Of General Layer Policies
+
+- It is the cross-profile behavior baseline that keeps execution consistent.
+- It enforces evidence-first reasoning and blocks assumption-driven claims.
+- It defines ticket-first execution so work is durable and resumable.
+- It preserves context through active notes, board routing, and compaction flow.
+- It enforces explicit approval gates before meaningful execution.
+- It sets escalation discipline for scope/risk changes.
+- It keeps security posture strict, especially around secret handling.
+- It preserves process quality across handoffs and context resets.
+
+### 9) What Must Be Internalized From General
+
+- Durable execution memory:
+  active findings belong in ticket notes, not transient chat context.
+- Optional context packs:
+  when a lane needs a reusable reread bundle, link it through the ticket's
+  `Context Management` section and `context_management/context_board.md`.
+- Execution gating:
+  no implementation/validation without ticket + routing + notes coherence.
+- Controlled scope:
+  no silent expansions, no hidden multi-file drift.
+- Re-entry reliability:
+  compaction is expected; re-onboarding and state reconstruction are mandatory.
+- Professional collaboration posture:
+  direct, explicit, and conflict-aware communication with evidence.
+
+### 10) Enforcement Reminder
+
+General policy is not optional glue text.
+It is the operating baseline for all inheriting profiles.
+If behavior in lower layers conflicts with this layer, resolve the conflict
+explicitly instead of bypassing general rules.
+
+### 11) Why These Behaviors Exist
+
+- They prevent drift between intent, execution, and recorded state.
+- They force decisions to be evidence-backed instead of memory-backed.
+- They keep work resumable after interruption, handoff, or compaction.
+- They reduce silent failure modes caused by skipped gates.
+
+### 12) General Failure Patterns To Avoid
+
+- Acting before scope and approval are explicit.
+- Continuing investigation without recording the current meaningful finding.
+- Treating board state and ticket state as optional or interchangeable.
+- Promoting weak assumptions into facts under time pressure.
+- Allowing completion state to diverge from actual acceptance.
+
+### 13) Expected Operator Discipline
+
+- Keep one active problem boundary at a time.
+- Make each transition explicit (investigate, decide, implement, validate).
+- Keep uncertainty visible until it is resolved with evidence.
+- Surface blockers early with concrete impact and next-step asks.
+- Preserve a clean audit trail through notes and state transitions.
+- Every assistant response must include one unique `prompt_id`.
+  - `prompt_id` is a per-response identifier, not a per-conversation constant.
+  - Use a fresh random lowercase hex/string value every response.
+  - Do not intentionally repeat a `prompt_id` within the same conversation.
+
+### 14) Continuity Standard
+
+General policy is designed so another reader can restart execution from files
+without replaying prior chat history.
+If a future reader cannot resume safely from ticket + board + notes, the
+current state is incomplete and must be repaired before continuing.
+
+
+
+
+
+
+
+

contextcompass_cli/_package/agent_onboarding/default/general/README.md

@@ -0,0 +1,38 @@
+
+
+# General Career (Shared)
+
+Purpose
+- Shared onboarding baseline for all careers.
+- Defines where execution policy, narrative workflow, and examples live.
+
+Primary entrypoints
+- `agent_onboarding/default/general/SKILLS.MD`: canonical read order.
+- `agent_onboarding/default/general/policies/policy_skills.md`: policy chain and execution gates.
+
+Folder map
+- `behavioral_guidelines/`: descriptive execution stories.
+  - `onboarding_summary.md`
+  - `agent_lifecycle_and_heartbeat.md`
+  - `work_intake_and_execution.md`
+- `policies/`: operational policy modules.
+- `skills/`: enforceable behavior and certification docs.
+- `examples/`: copy-first examples for ticket/doc structure references.
+  - `context_compass/examples/`
+
+Override rule
+- Career-specific docs should contain only true deltas from this shared baseline.
+- If content is shared across careers, keep it here instead of duplicating.
+
+Unknowns Gate
+- Apply the canonical policy in
+  `agent_onboarding/default/general/skills/unknowns_gate_reference.md`.
+- Local onboarding guidance must treat unevidenced claims as UNKNOWN.
+- Onboarding summaries must include concrete source pointers for promoted FACT claims.
+
+
+
+
+
+
+

contextcompass_cli/_package/agent_onboarding/default/general/SKILLS.MD

@@ -0,0 +1,92 @@
+
+
+# SKILLS.MD - general
+
+Purpose
+- Define active baseline skills for the `general` role.
+
+Inheritance
+- `INHERITS_SKILLS_FROM: none`
+
+Activation rule
+- All paths listed below are active skills for this role.
+- YOU MUST READ ALL PATHS BELOW for all these skills to be considered active.
+
+Active skills
+- `agent_onboarding/default/general/AGENTS.MD`
+- `agent_onboarding/default/general/skills/general.md`
+- `agent_onboarding/default/general/skills/execution_contract.md`
+- `agent_onboarding/default/general/skills/compaction_requirements.md`
+- `agent_onboarding/default/general/policies/policy_skills.md`
+- `agent_onboarding/default/general/skills/agent_stance.md`
+- `agent_onboarding/default/general/skills/agent_lifecycle.md`
+- `agent_onboarding/default/general/skills/agent_identity.md`
+- `agent_onboarding/default/general/skills/career_selection.md`
+- `agent_onboarding/default/general/skills/mrp_policy.md`
+- `agent_onboarding/default/general/skills/configuration_standards.md`
+- `agent_onboarding/default/general/skills/package_upgrade.md`
+- `agent_onboarding/default/general/skills/context_gold.md`
+- `agent_onboarding/default/general/skills/security_and_secrets.md`
+- `agent_onboarding/default/general/skills/repo_topology.md`
+- `agent_onboarding/default/general/skills/ticketing.md`
+- `agent_onboarding/default/general/skills/ticketing_skill_contract.md`
+- `agent_onboarding/default/general/skills/memory_management.md`
+- `agent_onboarding/default/general/skills/reactive_documentation.md`
+- `agent_onboarding/default/general/skills/active_documentation.md`
+- `agent_onboarding/default/general/skills/context_window_budget.md`
+- `agent_onboarding/default/general/skills/active_pointerboard.md`
+- `agent_onboarding/default/general/skills/mailbox_protocol.md`
+- `agent_onboarding/default/general/skills/ticket_closure_attention_sync.md`
+- `agent_onboarding/default/general/skills/unknowns_gate_reference.md`
+- `agent_onboarding/default/general/skills/self_certification.md`
+- `agent_onboarding/default/general/skills/user_approved_certification.md`
+- `agent_onboarding/default/general/skills/ticket_microcycle.md`
+- `agent_onboarding/default/general/skills/workflow.md`
+- `agent_onboarding/default/general/skills/role_local_workflows.md`
+- `agent_onboarding/default/general/WORKFLOWS.MD`
+- `agent_onboarding/default/general/workflows/cleanup_context_compass.md`
+- `agent_onboarding/default/general/workflows/start_context_compass_work.md`
+- `agent_onboarding/default/general/workflows/turn_in_selected_tickets.md`
+- `agent_onboarding/default/general/workflows/sync_attention_board.md`
+- `agent_onboarding/default/general/skills/context_compaction.md`
+- `attention_board.md`
+- `artifact_board.md`
+- `agent_onboarding/default/general/skills/context_management.md`
+- `context_management/context_board.md`
+
+Note on the three board paths above. `attention_board.md`, `artifact_board.md` and
+`context_management/context_board.md` are baseline reads, but they are **live
+mutable state, not policy**. You open them to learn what is happening right now, not
+to learn a rule. They therefore satisfy `READ_INTEGRITY_PROOF` by naming current
+state rather than a rule callout - see the live-state clause in
+`agent_onboarding/default/general/skills/compaction_requirements.md`. Do not invent
+a rule to quote from a routing table.
+- `agent_onboarding/default/general/behavioral_guidelines/agent_lifecycle_and_heartbeat.md`
+- `agent_onboarding/default/general/behavioral_guidelines/work_intake_and_execution.md`
+- `agent_onboarding/default/general/behavioral_guidelines/onboarding_summary.md`
+
+On-demand skills
+Read only when the trigger is met. Not part of the baseline, and claiming one
+was read without meeting its trigger is a false attestation.
+
+**This applies to sections labelled `On-demand`, which is the trigger-gated state.
+It is not a rule about everything outside baseline.** A `Self-directed` section is
+also outside baseline but has no trigger and needs no permission - waiting for one
+there is the failure, not the discipline. `engineer` carries such a section for its
+system documents. Three states, defined in `context_compass/SKILLS.MD`.
+
+- `agent_onboarding/default/general/skills/llm_full_usage.md`
+  - Trigger: `llm_full.md` exists at the repository root and you are about to
+    read from it, or you have been asked to regenerate it.
+  - Why it is on-demand: most installs have no `llm_full.md` - it is a view of
+    the package produced for whole-system review, not something the package
+    ships. Reading the skill when the file is absent teaches a procedure with
+    nothing to apply it to.
+  - Why it is not optional once triggered: the document is ~27,000 lines. The
+    skill exists to stop you reading it whole when a 450-line index would have
+    told you the exact range.
+
+
+
+
+

contextcompass_cli/_package/agent_onboarding/default/general/WORKFLOWS.MD

@@ -0,0 +1,23 @@
+# WORKFLOWS.MD - general
+
+Purpose
+- Define role-local workflow availability for the `general` role.
+
+Inheritance
+- `INHERITS_WORKFLOWS_FROM: none`
+
+Activation rule
+- Workflows are optional macro behaviors.
+- Workflows are user-generated and user-approved only.
+- When a workflow is selected for work, treat the parent-first role chain as
+  the workflow lookup order.
+
+Active workflows
+- `agent_onboarding/default/general/workflows/cleanup_context_compass.md`
+- `agent_onboarding/default/general/workflows/start_context_compass_work.md`
+- `agent_onboarding/default/general/workflows/turn_in_selected_tickets.md`
+- `agent_onboarding/default/general/workflows/sync_attention_board.md`
+
+On-demand workflows
+- `agent_onboarding/default/general/workflows/role_creation.md`
+- `agent_onboarding/default/general/workflows/workflow_creation.md`

contextcompass_cli/_package/agent_onboarding/default/general/behavioral_guidelines/README.md

@@ -0,0 +1,12 @@
+# Behavioral Guidelines Index
+
+Purpose
+- Provide a quick index for general behavioral guidance docs.
+
+Primary guides
+- `work_intake_and_execution.md`
+- `onboarding_summary.md`
+- `agent_lifecycle_and_heartbeat.md`
+
+Use these as behavior deltas.
+Policy and routing authority remain in entrypoint, config, and `SKILLS.MD` chain.

contextcompass_cli/_package/agent_onboarding/default/general/behavioral_guidelines/agent_lifecycle_and_heartbeat.md

@@ -0,0 +1,37 @@
+
+
+# agent_lifecycle_and_heartbeat
+
+Purpose
+- Define the session lifecycle and handoff habits for this repo.
+
+Story steps
+1) Establish session identity
+   - Use the agent identity provided by the user.
+   - If identity is missing or unclear, stop and ask before any work.
+2) Check-in
+   - Acknowledge scope, constraints, and ticket status.
+   - Ensure `attention_board.md` reflects active routing and open the linked active ticket.
+3) Work loop
+   - Use `tickets/epics/`, `tickets/stories/`, and `tickets/tasks/` for planning and execution.
+   - Keep ticket checklists current and append findings as they occur in `## Notes`.
+4) Check-out
+   - Summarize what changed, what remains, and what to do next.
+   - Move completed tickets to their matching completed folder
+     (`tickets/epics/completed/`, `tickets/stories/completed/`, `tickets/tasks/completed/`) only after user confirmation.
+
+Artifacts touched
+- `attention_board.md`
+- `tickets/epics/`
+- `tickets/epics/completed/`
+- `tickets/stories/`
+- `tickets/stories/completed/`
+- `tickets/tasks/`
+- `tickets/tasks/completed/`
+
+References
+- `agent_onboarding/default/general/skills/agent_lifecycle.md`
+- `AGENTS.MD`
+
+
+

contextcompass_cli/_package/agent_onboarding/default/general/behavioral_guidelines/onboarding_summary.md

@@ -0,0 +1,41 @@
+
+
+# onboarding_summary
+
+Purpose
+- Provide a concise onboarding checklist for agents.
+
+Checklist (short form)
+1) Confirm repo root + policy sources
+   - Read `AGENTS.MD` and any directory-local `AGENTS.MD` in scope.
+2) Read onboarding entrypoints
+   - `config/context_compass_config.yaml`
+   - `context_compass/SKILLS.MD`
+3) First-time user path only (`new` profile)
+   - Read `agent_onboarding/default/new/skills/first_time_profile_setup.md`
+   - Read `agent_onboarding/default/new/README.md`
+4) Select working role
+   - Read the registry table in `context_compass/SKILLS.MD` for the current
+     role list. Do not rely on a role list copied into any other document.
+   - Read `agent_onboarding/default/general/SKILLS.MD` first.
+   - Read the selected role `SKILLS.MD` chain after baseline, parent-first.
+5) Certification gate
+   - Request approval before any tool usage or edits.
+   - Approval must include the exact token `CERTIFY: APPROVED`.
+6) Post-cert work execution
+   - Use `tickets/epics/`, `tickets/stories/`, and `tickets/tasks/` for all
+     work.
+   - Route from `attention_board.md` and resume from linked ticket notes.
+   - Orient from `system_docs/src_architecture.md` plus the architecture and
+     component indexes, then descend: index -> slice -> the code. Slice
+     `src_components.md` and `src_graph.md` through their indexes whenever the
+     work needs them - your call, no permission required, and never whole.
+   - Keep `attention_board.md` current for routing and ticket notes current for
+     durable findings.
+
+References
+- `AGENTS.MD`
+- `config/context_compass_config.yaml`
+- `context_compass/SKILLS.MD`
+
+

contextcompass_cli/_package/agent_onboarding/default/general/behavioral_guidelines/work_intake_and_execution.md

@@ -0,0 +1,44 @@
+
+
+# work_intake_and_execution
+
+Purpose
+- Define how we intake work and execute it using `tickets/epics/`, `tickets/stories/`, and `tickets/tasks/` tickets.
+
+Story steps
+1) Capture intent
+   - Restate the goal, constraints, and scope.
+   - If unclear, ask before writing tickets.
+2) Create tickets
+   - Use `templates/` to create epic/story/task tickets.
+   - Keep tickets deep and descriptive for handoff durability.
+3) Confirm before edits
+   - Share the plan and files to touch.
+   - Wait for explicit approval before edits.
+4) Execute
+   - Follow ticket steps and update checklists as work progresses.
+   - Run the Ticket Microcycle and document each meaningful finding before continuing investigation.
+   - Keep changes scoped and reviewable.
+5) Close out
+   - Walk through what was delivered.
+   - Ask if acceptance criteria are met.
+   - Move tickets to their matching completed folder (`tickets/epics/completed/`,
+     `tickets/stories/completed/`, `tickets/tasks/completed/`) only after explicit confirmation.
+
+Artifacts touched
+- `templates/`
+- `tickets/epics/`
+- `tickets/epics/completed/`
+- `tickets/stories/`
+- `tickets/stories/completed/`
+- `tickets/tasks/`
+- `tickets/tasks/completed/`
+
+References
+- `SKILLS.MD`
+- `agent_onboarding/default/general/skills/workflow.md`
+
+
+
+
+