specify

User Input

You MUST consider the user input before proceeding (if not empty).

Outline

The text the user typed after /akka:specify in the triggering message is the feature description. Assume you always have it available in this conversation even if the user's request appears literally below. Do not ask the user to repeat it unless they provided an empty command.

Given that feature description, do this:

Prerequisite check: Before doing anything else, verify the environment is ready:
- Call akka_sdd_list_specs — if this MCP tool is not available, the Akka MCP server is not configured. Stop and tell the user: "The Akka MCP server is not running. Run /akka:setup first to configure your project, or run akka specify init . from the terminal."
- Check if the current directory has a pom.xml or .akka/ directory. If neither exists, warn: "This doesn't look like an Akka project directory. Run /akka:setup to scaffold a new project, or cd into an existing one."
- If both checks pass, proceed.
Generate a concise short name (2-4 words) for the branch:
- Analyze the feature description and extract the most meaningful keywords
- Create a 2-4 word short name that captures the essence of the feature
- Use action-noun format when possible (e.g., "add-user-auth", "fix-payment-bug")
- Preserve technical terms and acronyms (OAuth2, API, JWT, etc.)
- Keep it concise but descriptive enough to understand the feature at a glance
- Examples:
  - "I want to add user authentication" → "user-auth"
  - "Implement OAuth2 integration for the API" → "oauth2-api-integration"
  - "Create a dashboard for analytics" → "analytics-dashboard"
  - "Fix payment processing timeout bug" → "fix-payment-timeout"
Check for existing branches before creating new one:

a. First, fetch all remote branches to ensure we have the latest information.

b. Find the highest feature number across all sources for the short-name:
- Remote branches: git ls-remote --heads origin | grep -E 'refs/heads/[0-9]+-<short-name>$'
- Local branches: git branch | grep -E '^[* ]*[0-9]+-<short-name>$'
- Specs directories: Check for directories matching specs/[0-9]+-<short-name>
c. Determine the next available number:
- Extract all numbers from all three sources
- Find the highest number N
- Use N+1 for the new branch number
d. Call the akka_sdd_create_spec MCP tool with feature_name set to a short, hyphenated name derived from the user's feature description (e.g. "user-authentication"). The tool creates a numbered feature directory under the specs directory, initializes spec.md from the template, and returns the path. Use this path as FEATURE_SPEC and its parent directory as FEATURE_DIR.
- Pass number (N+1) and short_name parameters along with the feature description
IMPORTANT:
- Check all three sources (remote branches, local branches, specs directories) to find the highest number
- Only match branches/directories with the exact short-name pattern
- If no existing branches/directories found with this short-name, start with number 1
- The tool output will contain BRANCH_NAME and SPEC_FILE paths
Load akka_sdd_get_template to understand required sections.
Follow this execution flow:
1. Parse user description from Input If empty: ERROR "No feature description provided"
2. Extract key concepts from description Identify: actors, actions, data, constraints For agent features, apply the conditional "Agent behavior, protection, and evaluation requirements" guidance below.
3. For unclear aspects:
  - Make informed guesses based on context and industry standards
  - Only mark with [NEEDS CLARIFICATION: specific question] if:
    - The choice significantly impacts feature scope or user experience
    - Multiple reasonable interpretations exist with different implications
    - No reasonable default exists
  - LIMIT: Maximum 3 [NEEDS CLARIFICATION] markers for routine details; policy decisions always keep a marker and do not count toward the limit (see "Policy decisions are not reasonable defaults")
  - Prioritize clarifications by impact: scope > security/privacy > user experience > technical details
4. Fill User Scenarios & Testing section If no clear user flow: ERROR "Cannot determine user scenarios"
5. Generate Functional Requirements Each requirement must be testable Use reasonable defaults for unspecified details (document assumptions in Assumptions section)
6. Define Success Criteria Create measurable, technology-agnostic outcomes Include both quantitative metrics (time, performance, volume) and qualitative measures (user satisfaction, task completion) Each criterion must be verifiable without implementation details
7. Identify Key Entities (if data involved)
8. Return: SUCCESS (spec ready for planning)
Write the specification to SPEC_FILE using the template structure, replacing placeholders with concrete details derived from the feature description (arguments) while preserving section order and headings.
Specification Quality Validation: After writing the initial spec, validate it against quality criteria:

a. Create Spec Quality Checklist: Generate a checklist file at FEATURE_DIR/checklists/requirements.md using the checklist template structure with these validation items:
```
# Specification Quality Checklist: [FEATURE NAME]

**Purpose**: Validate specification completeness and quality before proceeding to planning
**Created**: [DATE]
**Feature**: [Link to spec.md]

## Content Quality

- [ ] No implementation details (languages, frameworks, APIs)
- [ ] Focused on user value and business needs
- [ ] Written for non-technical stakeholders
- [ ] All mandatory sections completed

## Requirement Completeness

- [ ] No [NEEDS CLARIFICATION] markers remain
- [ ] Requirements are testable and unambiguous
- [ ] Success criteria are measurable
- [ ] Success criteria are technology-agnostic (no implementation details)
- [ ] All acceptance scenarios are defined
- [ ] Edge cases are identified
- [ ] Scope is clearly bounded
- [ ] Dependencies and assumptions identified

## Feature Readiness

- [ ] All functional requirements have clear acceptance criteria
- [ ] User scenarios cover primary flows
- [ ] Feature meets measurable outcomes defined in Success Criteria
- [ ] No implementation details leak into specification

## Notes

- Items marked incomplete require spec updates before `/akka:clarify` or `/akka:plan`
```
b. Run Validation Check: Review the spec against each checklist item:
- For each item, determine if it passes or fails
- Document specific issues found (quote relevant spec sections)
c. Handle Validation Results:
- If all items pass: Mark checklist complete and proceed to step 6
- If items fail (excluding [NEEDS CLARIFICATION]):
  1. List the failing items and specific issues
  2. Update the spec to address each issue
  3. Re-run validation until all items pass (max 3 iterations)
  4. If still failing after 3 iterations, document remaining issues in checklist notes and warn user
- If [NEEDS CLARIFICATION] markers remain:
  1. Extract all [NEEDS CLARIFICATION: ...] markers from the spec
  2. LIMIT CHECK: If more than 3 non-policy markers exist, keep only the 3 most critical (by scope/security/UX impact) and make informed guesses for the rest. Never guess policy decisions: they keep their markers, and the spec is not ready until they are resolved
  3. For each clarification needed (max 3), present options to user in this format:
```
## Question [N]: [Topic]

**Context**: [Quote relevant spec section]

**What we need to know**: [Specific question from NEEDS CLARIFICATION marker]

**Suggested Answers**:

| Option | Answer | Implications |
|--------|--------|--------------|
| A      | [First suggested answer] | [What this means for the feature] |
| B      | [Second suggested answer] | [What this means for the feature] |
| C      | [Third suggested answer] | [What this means for the feature] |
| Custom | Provide your own answer | [Explain how to provide custom input] |

**Your choice**: _[Wait for user response]_
```
  4. CRITICAL - Table Formatting: Ensure markdown tables are properly formatted:
    - Use consistent spacing with pipes aligned
    - Each cell should have spaces around content: | Content | not |Content|
    - Header separator must have at least 3 dashes: |--------|
    - Test that the table renders correctly in markdown preview
  5. Number questions sequentially (Q1, Q2, Q3 - max 3 total)
  6. Present all questions together before waiting for responses
  7. Wait for user to respond with their choices for all questions (e.g., "Q1: A, Q2: Custom - [details], Q3: B")
  8. Update the spec by replacing each [NEEDS CLARIFICATION] marker with the user's selected or provided answer
  9. Re-run validation after all clarifications are resolved
d. Update Checklist: After each validation iteration, update the checklist file with current pass/fail status
Report completion with branch name, spec file path, checklist results, and readiness for the next phase (/akka:clarify or /akka:plan).

NOTE: The script creates and checks out the new branch and initializes the spec file before writing.

General Guidelines

Quick Guidelines

Focus on WHAT users need and WHY.
Avoid HOW to implement (no tech stack, APIs, code structure).
Written for business stakeholders, not developers.
DO NOT create any checklists that are embedded in the spec. That will be a separate command.

Section Requirements

Mandatory sections: Must be completed for every feature
Optional sections: Include only when relevant to the feature
When a section doesn't apply, remove it entirely (don't leave as "N/A")

For AI Generation

When creating this spec from a user prompt:

Make informed guesses: Use context, industry standards, and common patterns to fill gaps
Document assumptions: Record reasonable defaults in the Assumptions section
Limit clarifications: Maximum 3 [NEEDS CLARIFICATION] markers for routine details, not counting policy decisions, which are never guessed - use only for critical decisions that:
- Significantly impact feature scope or user experience
- Have multiple reasonable interpretations with different implications
- Lack any reasonable default
Prioritize clarifications: scope > security/privacy > user experience > technical details
Think like a tester: Every vague requirement should fail the "testable and unambiguous" checklist item
Common areas needing clarification (only if no reasonable default exists):
- Feature scope and boundaries (include/exclude specific use cases)
- User types and roles (if multiple conflicting interpretations possible)
- Security/compliance requirements (when legally/financially significant)

Examples of reasonable defaults (don't ask about these):

Performance targets: Standard web/mobile app expectations unless specified
Error handling: User-friendly messages with appropriate fallbacks
Authentication method: Standard session-based or OAuth2 for web apps
Integration patterns: Use project-appropriate patterns (REST/GraphQL for web services, function calls for libraries, CLI args for tools, etc.)

Policy decisions are not reasonable defaults. Do not invent consequential thresholds, access permissions, retention periods, sensitive-data rules, or approval requirements. Reuse established requirements when available; otherwise mark each with [NEEDS CLARIFICATION]. These markers are not capped at 3: ask up to 3 here and leave the rest for /akka:clarify. A spec with open policy markers is not ready for /akka:plan.

Agent behavior, protection, and evaluation requirements

Apply this guidance only when the feature involves agents or model-generated behavior. Use the description, existing service requirements, policies, and configuration to answer these questions first; ask only where a missing decision affects scope or acceptance. Do not add controls or a governance workflow merely because an agent exists.

Enforcement boundaries: Identify what must be checked before sending input to a model, executing a tool, or delivering a reply. State whether a violation blocks the operation, transforms data, or is recorded for review. Define behavior when the check fails to run or cannot reach a decision. Where a classifier informs the check, define the meaning of labels or scores, the decision criteria, and handling of uncertainty, invalid output, or classifier failure; distinguish the classification signal from the action taken on it.
Streaming: Establish whether replies must stream immediately or pass a complete-response check before delivery. Surface conflicting requirements: a check after delivery cannot prevent disclosure of content already streamed.
Sensitive data: Identify the data to protect, the agents in scope, and the relevant destinations: model input, tool results, logs, or external responses. Specify the required transformation or restriction for each; do not assume one sanitizer covers every boundary or historical records.
Retrospective evaluation: Define the quality criteria, which interactions to assess, whether failed interactions are included, the required coverage or sampling, and the expected response to adverse findings. An assessment after completion cannot prevent the action it assesses. Define which decisions and assessments must be traceable, with the identifiers, policy/model versions, and outcomes needed for investigation, subject to the data-handling requirements.
Access control: Establish which users, services, and components may invoke each operation. Distinguish component identity, end-user permissions, and tool authorization, including resource and purpose restrictions where required.

Keep requirements focused on observable behavior and free of Akka SDK interfaces or component names. Record constraints stated by the user (existing implementations, required Akka SDK version) under Assumptions (add the section if the template lacks one); /akka:plan selects the mechanisms.

Success Criteria Guidelines

Success criteria must be:

Measurable: Include specific metrics (time, percentage, count, rate)
Technology-agnostic: No mention of frameworks, languages, databases, or tools
User-focused: Describe outcomes from user/business perspective, not system internals
Verifiable: Can be tested/validated without knowing implementation details

For probabilistic model behavior, define acceptance over an agreed, representative dataset with repeated trials, measurable error limits, and a stated aggregation rule. Include adverse and uncertain cases where relevant. Use stakeholder-approved targets; one passing example or a model's self-reported confidence does not establish quality. Leave the choice of verification tooling to Plan.

Good examples:

"Users can complete checkout in under 3 minutes"
"System supports 10,000 concurrent users"
"95% of searches return results in under 1 second"
"Task completion rate improves by 40%"
"Across the 200-case approved support dataset, run 5 times each, fewer than 1% of replies are harmful and at least 95% of benign requests are answered"

Bad examples (implementation-focused):

"API response time is under 200ms" (too technical, use "Users see results instantly")
"Database can handle 1000 TPS" (implementation detail, use user-facing metric)
"React components render efficiently" (framework-specific)
"Redis cache hit rate above 80%" (technology-specific)

Exit conditions and the feedback loop

Akka Specify turns requirements into machine-checked exit conditions, each backed by a check that delegates to the ecosystem's own tooling. In Enforced mode the engine will not let the build advance until every check passes and has been reviewed. In À la carte mode every check is advisory — the user always sees the full picture and the run never blocks on a failing condition.

Check first whether this loop applies at all. Run akka specify mode. In À la carte mode the exit-condition set is dormant until the developer asks for it, and the whole of this section is then skipped: capture nothing, create no auditors, call no akka_ec_* tool, and do not report a verdict. Treat /akka:specify as spec-and-plan work and nothing more. The dormant state is not a gap to close and not something to talk the user into — mention exit conditions only if they ask what else is available, and never more than once. If the developer does ask for checks, run akka specify mode a-la-carte --exit-conditions=honor and then follow this loop with every result advisory.

When the set is active, on each /akka:specify <feedback> turn:

1. Capture feedback as exit conditions

For each requirement the feedback expresses:

Classify — an answer to an open decision (set the value in an existing condition), a new checkable invariant (a new exit condition), a revealed-inadequate check (tighten it), or a one-off/subjective item (do the task, or ask — do not invent a check).
Falsifiability gate — only create an exit condition if you can state a binary pass predicate that is FALSE exactly when the requirement is violated, over an observable signal. If you cannot, ask a clarifying question instead of inventing a meaningless check.
Define the condition — id, tier, dod_type, a one-line invariant, and how it is checked.
Choose the check by delegation — prefer the tool that already decides this (tsc --noEmit, mvn compile/test, go build, a linter). Author it as a declarative command: {run: [...], applies_to: {requirement, module}}. Write bespoke logic only when no tool observes the invariant.
Must not — do not create a check that always passes, that could pass while the invariant is false, that is subjective without a probe, that is over-broad, or that duplicates an existing condition.
Record it with akka_ec_capture (the condition as JSON). The check will not run until the human approves it.

2. Surface the exit conditions for the human, at the exit-condition level

Present the new/changed exit conditions in plain language — "I'd like to add these exit conditions," each with a one-line description and how I'll check it — and ask the human to approve / adjust / skip. Never phrase this as "approve running a command"; it is always about the checks. When the human approves, call akka_ec_approve with the condition id.

3. Review the checks adversarially

Before finishing a turn (and always before ship), for each check attempt to construct an input where the check passes while its invariant is false. If you can, the check is inadequate — fix it or ask. Then call akka_ec_adequacy_submit to record the review for the current checks.

4. Report plainly

After any feedback, call akka_ec_conform, then present akka_ec_summary to the user verbatim — it is the plain outcome (decisions needed / building / ready). Do NOT restate internal condition ids, check keys, or the words "auditor" / "coverage gate" / "adequacy review". Offer "say 'show details' for the full checklist" for the manifest.

Routing and mode

Once a build is active, treat every reply as feedback and run this loop — the user does not need to prefix /akka:specify each time. The gates are state-based, so a reply that is not recorded simply does not advance the build and the pending item re-surfaces (a stall, never a false pass). This capture-and-review logic is identical wherever the exit-condition set is active; Enforced mode additionally blocks on the gates. Where the set is dormant the loop does not run at all, so there are no gates to stall on.

Companion generators

Two surfaces beyond code have their own generator commands, each paired with an exit-condition family the engine verifies:

/akka:harnesses generates the enterprise-configuration assets (CI, scanning, supply-chain, content style packages) into /harnesses/ and records them in .akka/harnesses.lock. Its family is activated / configured / attested.
/akka:docs generates the rendered documentation under docs/. Its family is language, structure, completeness, and tone.

Generation runs in the assistant; the engine detects the surfaces, verifies deterministically, and records the attestations the assistant submits. Both follow the same mode invariant — advisory where the exit-condition set is active in À la carte, blocking in Enforced — so a coverage gap in a harness or a page reds the same single gate as an uncovered code module. Both generators still run while the set is dormant; they simply produce no conditions and no gate.

Done When

The mode was read before any akka_ec_* call, and where the exit-condition set was dormant nothing was captured, approved, audited, or reported as a verdict.
Where the set was active: any checkable requirement in the feedback was captured as an exit condition + check via akka_ec_capture, surfaced for approval at the exit-condition level, and approved via akka_ec_approve.
Where the set was active: the active checks were reviewed adversarially and akka_ec_adequacy_submit was called; the plain akka_ec_summary (never internal ids) was shown to the user.
A short-name and next feature number were derived by checking remote branches, local branches, and the specs/ directory; the highest existing N was used and N+1 was passed to akka_sdd_create_spec.
akka_sdd_create_spec created a numbered FEATURE_DIR and returned BRANCH_NAME and SPEC_FILE.
The specification was written to SPEC_FILE using the template structure, with all mandatory sections filled and section order preserved.
Success Criteria are measurable, technology-agnostic, user-focused, and verifiable.
No more than 3 [NEEDS CLARIFICATION] markers for routine details remain (surplus resolved by informed guesses documented in Assumptions); no policy decision was guessed, and open ones remain as markers.
FEATURE_DIR/checklists/requirements.md was created and every non-[NEEDS CLARIFICATION] item passes (or the user was warned after 3 failed iterations).
Completion was reported to the user with branch name, spec file path, checklist results, and the next-command recommendation (/akka:clarify or /akka:plan).