Deployment Manifest: {Product Name}

Order: {order-id} Client: {client-name} Product type: {product-type} Deployment Specialist: Diego Date: {date}

1. Product Information

Field	Value
Product name	{product-name}
Order ID	{order-id}
Client	{client-name}
Product type	{product-type}
Version	{version}
QA certification	{PASS/CONDITIONAL} -- certified by Quinn on {qa-date}
Working language	{language}

Field	Value
Selected option	{A / B / C}
Option description	{Description -- see below}

Option	Method	When to Use
A	New repository -- client receives a fresh Git repository	New client, first product
B	Branch/PR into existing client repository	Existing client, adding a new talent alongside others
C	Update to existing talent repository	Enhancement or update to a deployed talent (Type 3 request)

{Complete the section matching the selected option. Delete the others.}

Option A -- New Repository:

Option B -- Existing Repository:

Option C -- Update:

All files and directories included in or excluded from the deployment package.

#	File / Directory	Include / Exclude	Reason
1	`CLAUDE.md`	Include	Core agent configuration
2	`README.md`	Include	Client-facing overview
3	`.claude/commands/*.md`	Include	All product skills
4	`.claude/commands/templates/`	Include	Output templates
5	`.claude/settings.json`	Include	Shipping permission configuration
6	`.claude/settings.local.json`	Exclude	Development-only settings
7	`content-in/`	Include	Reference materials
8	`workflows/`	Include / Exclude	{Include if orchestration, exclude if not used}
9	`memory/`	Include	Persistent memory directory (empty or with seed data)
10	`.gitignore`	Include	Git configuration
11	`docs/user-guide.md`	Include	Client documentation
12	`docs/config-guide.md`	Include	Client documentation
13	`docs/skill-reference-card.md`	Include	Client documentation
14	`docs/quick-start.md`	Include	Client documentation
{n}	`{path}`	{Include/Exclude}	{Reason}

Total files included: __ Total files excluded: __

Before packaging, verify all development artifacts are removed or excluded.

No development-only settings -- .claude/settings.local.json excluded from package
No internal notes or TODOs -- All internal comments, TODO files, and architecture analysis documents removed
No test data in output directories -- Output directories are clean or contain only seed/example data explicitly intended for the client
No factory references -- No references to Talent Factory internal processes, order IDs, or factory worker names in client-facing files
No sensitive credentials -- No API keys, tokens, passwords, or internal URLs in any shipped files

Sanitization result: PASS / FAIL Issues found: {list any issues, or "None"}

Deployment verification checks matching Gate 6 criteria.

#	Check	Method	Expected Result	Actual Result	Status
1	Agent recognizes its role	Ask "Who are you?"	Responds with correct role and client name	{Actual}	PASS/FAIL
2	Agent responds in correct language	Ask a question in {language}	Responds entirely in {language}	{Actual}	PASS/FAIL
3	All skills invocable	Run `/help` and invoke each skill	All {count} skills listed and executable	{Actual}	PASS/FAIL
4	Reference materials accessible	Ask about content in `content-in/`	Agent references loaded materials correctly	{Actual}	PASS/FAIL
5	Client-specific settings active	Check permissions and conventions	Settings match `.claude/settings.json` and CLAUDE.md	{Actual}	PASS/FAIL
6	Output directory writable	Run a skill that produces output	Output file created in correct location	{Actual}	PASS/FAIL
{n}	{Additional check}	{Method}	{Expected}	{Actual}	PASS/FAIL

Verification checks passed: __ / __ total Verification result: PASS / FAIL

In case the deployment needs to be reversed:

Rollback is initiated if:

{Trigger 1 -- e.g., "Client reports critical skill failures within 48 hours of deployment"}
{Trigger 2 -- e.g., "Verification checks fail in client environment"}
{Additional triggers}

{Step 1 -- e.g., "Notify client that rollback is in progress"}
{Step 2 -- e.g., "Revert to previous version: git revert {commit-hash}" or "Remove repository access"}
{Step 3 -- e.g., "Confirm with client that previous state is restored"}
{Step 4 -- e.g., "Open remediation ticket and route to Pablo for rebuild"}