Product
Create or refine a product contract with the user's authority.
The contract works because it forces every claim into evidence or labeled
aspiration; if a reader cannot tell which is which, the document has failed.
- Inspect existing product, README, goals, release, and evidence sources.
- Ask only for decisions that cannot be grounded safely from those sources.
- State mission, users, pains, value, differentiation, and non-goals. For
this repository's own product surfaces, start from the canonical category
in
docs/contracts/ubiquitous-language.md (AgentOps is the operations
layer for agentic engineering) and preserve the small ownership boundary —
never re-frame AgentOps as an execution orchestrator, factory, or loop.
- Separate evidence from hope under two required headings. Put every grounded
claim under
## Proven, each carrying a resolvable citation (a README,
release, test, or evidence source). Put unproven hopes, evidence gaps, and
not-yet-measured success signals under ## Assumptions. A claim that cannot
cite a source belongs under ## Assumptions, never ## Proven.
- Preserve an existing PRODUCT.md unless the user authorizes replacement.
- Return the document to the caller; Plan may use it as intent context.
Named failure mode — aspiration laundering: an unproven hope written in the
## Proven section; once laundered, every downstream plan inherits a false
premise. Detector: any ## Proven claim without a resolvable citation is a
laundered aspiration — move it to ## Assumptions or cite it.
Anti-pattern: rewriting a healthy PRODUCT.md wholesale because the session has
fresh opinions. Corrective: refine only the sections the user asked to change
and preserve the rest byte-for-byte.
Product does not select work, invoke a loop, repair itself, validate itself, or
choose delivery.
1---2name: product3description: Create or refine PRODUCT.md while separating evidence, aspiration, users, value, and non-goals. Triggers: "product", "create PRODUCT.md", "product boundary".4---5
6# Product
7
8Create or refine a product contract with the user's authority.
9
10The contract works because it forces every claim into evidence or labeled
11aspiration; if a reader cannot tell which is which, the document has failed.
12
131. Inspect existing product, README, goals, release, and evidence sources.
142. Ask only for decisions that cannot be grounded safely from those sources.
153. State mission, users, pains, value, differentiation, and non-goals. For
16 this repository's own product surfaces, start from the canonical category
17 in `docs/contracts/ubiquitous-language.md` (AgentOps is the operations
18 layer for agentic engineering) and preserve the small ownership boundary —
19 never re-frame AgentOps as an execution orchestrator, factory, or loop.
204. Separate evidence from hope under two required headings. Put every grounded
21 claim under `## Proven`, each carrying a resolvable citation (a README,
22 release, test, or evidence source). Put unproven hopes, evidence gaps, and
23 not-yet-measured success signals under `## Assumptions`. A claim that cannot
24 cite a source belongs under `## Assumptions`, never `## Proven`.
255. Preserve an existing PRODUCT.md unless the user authorizes replacement.
266. Return the document to the caller; Plan may use it as intent context.
27
28Named failure mode — **aspiration laundering**: an unproven hope written in the
29`## Proven` section; once laundered, every downstream plan inherits a false
30premise. Detector: any `## Proven` claim without a resolvable citation is a
31laundered aspiration — move it to `## Assumptions` or cite it.
32
33Anti-pattern: rewriting a healthy PRODUCT.md wholesale because the session has
34fresh opinions. Corrective: refine only the sections the user asked to change
35and preserve the rest byte-for-byte.
36
37Product does not select work, invoke a loop, repair itself, validate itself, or
38choose delivery.