heimdall-agent 1.1.0 → 1.2.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (41) hide show
  1. package/README.md +13 -13
  2. package/README.pt-BR.md +36 -22
  3. package/agents/code-sentinel.md +7 -1
  4. package/agents/modernizer.md +4 -2
  5. package/agents/orchestrator.md +20 -7
  6. package/agents/reviewer.md +2 -0
  7. package/config/agents.json +11 -1
  8. package/config/mcp.json +1 -1
  9. package/config/profiles.json +1 -1
  10. package/config/skills.json +762 -683
  11. package/docs/SKILLS.md +96 -74
  12. package/docs/runs/skill-audit/2026-11-12-compare-report.json +1028 -0
  13. package/docs/runs/skill-audit/2026-11-12-upstream-audit.json +275 -0
  14. package/doctrine/AGENTS.md +4 -1
  15. package/lib/catalog-sync.cjs +8 -3
  16. package/package.json +1 -1
  17. package/playbooks/01-interactive-grilling-gate.md +14 -10
  18. package/playbooks/03-adversarial-code-review.md +12 -4
  19. package/playbooks/05-canonical-plan-template.md +3 -1
  20. package/setup.mjs +2 -2
  21. package/skills/antislop-copywriting/SKILL.md +36 -1
  22. package/skills/brainstorming/SKILL.md +60 -38
  23. package/skills/codebase-design/DEEPENING.md +37 -0
  24. package/skills/codebase-design/DESIGN-IT-TWICE.md +44 -0
  25. package/skills/codebase-design/SKILL.md +114 -0
  26. package/skills/domain-modeling/ADR-FORMAT.md +47 -0
  27. package/skills/domain-modeling/GLOSSARY-FORMAT.md +60 -0
  28. package/skills/domain-modeling/SKILL.md +74 -0
  29. package/skills/executing-plans/SKILL.md +6 -4
  30. package/skills/grill-me/SKILL.md +65 -0
  31. package/skills/handoff/SKILL.md +64 -0
  32. package/skills/improve-codebase-architecture/HTML-REPORT.md +123 -0
  33. package/skills/improve-codebase-architecture/SKILL.md +71 -0
  34. package/skills/owasp-security/reference/config-and-supply-chain.md +346 -344
  35. package/skills/owasp-security/reference/languages.md +378 -372
  36. package/skills/owasp-security/reference/owasp-report.md +1144 -1140
  37. package/skills/owasp-security/reference/review-checklist.md +134 -132
  38. package/skills/subagent-driven-development/SKILL.md +7 -4
  39. package/skills/to-spec/SKILL.md +75 -0
  40. package/skills/to-tickets/SKILL.md +105 -0
  41. package/skills/wayfinder/SKILL.md +128 -0
@@ -31,6 +31,10 @@ Before you start implementing, you must, in order:
31
31
  3. Get their approval of that doc.
32
32
  4. Invoke the superpowers:writing-plans skill.
33
33
 
34
+ If nothing is open when you write your first message and there's one
35
+ sensible way, steps 1 to 3 happen in that message (see Your First
36
+ Message).
37
+
34
38
  Two things aren't implementing: a spike they said yes to (see Spike to
35
39
  feel things out), and doing what they asked once they've opted out of
36
40
  the questions.
@@ -42,7 +46,8 @@ a handoff, files they pointed at, earlier conversation), dig in before your
42
46
  first message. Read what they referred to and the code or material it
43
47
  connects to, and keep reading until you can explain how the relevant
44
48
  parts work today and what this change would touch. Never ask what the
45
- project or their context already answers.
49
+ project or their context already answers, or reopen what they've
50
+ already decided.
46
51
 
47
52
  ## Your First Message
48
53
 
@@ -58,8 +63,17 @@ Move the work forward, with one of these:
58
63
  has to be built, like a constraint, a conflict, or a cheaper path, say
59
64
  so plainly. How to build it comes later, once you understand what they
60
65
  want (see Lay out the approaches). End with one line inviting
61
- corrections, then one question about what they want: the thing their
62
- request leaves open whose answer would most change what gets built.
66
+ corrections. Then, if they haven't told you why they want it, even if
67
+ you can guess, ask about that: one open question in their own terms,
68
+ with nothing to pick from. If they have told you, and their request
69
+ and the project leave open something whose answer would change what
70
+ gets built, ask about the one that would change it most; when its
71
+ choices can be named, sketch each concretely, with what it trades off
72
+ and which you'd pick. If nothing like that is open, there's one
73
+ sensible way: end with the short design itself (the approach, a line on
74
+ why the others don't fit, and the design in a few lines) and ask them
75
+ to approve it. Their approval covers the approach and the design: write
76
+ the short design doc as approved and go on to the plan.
63
77
  Specs are often incomplete, so confirm before you build on one.
64
78
  - **They haven't:** one open question, in their own terms, about the
65
79
  biggest thing their request leaves unsettled. Say why you're asking:
@@ -72,10 +86,9 @@ How the message ends depends on whether their message asked for
72
86
  brainstorming:
73
87
 
74
88
  - **They asked for it** (by name, by invoking it, or by asking for a
75
- design discussion): end with your question. They want this
76
- conversation.
77
- - **They didn't:** they may not want a conversation at all. After your
78
- question, add a light, one-line offer to just get going instead, like
89
+ design discussion): end there. They want this conversation.
90
+ - **They didn't:** they may not want a conversation at all. End with a
91
+ light, one-line offer to just get going instead, like
79
92
  "Or if you'd rather I just dive in, say so." If they take you up on
80
93
  it, this skill is done: state any significant how-it-gets-made choice
81
94
  (platform, language, medium) in one plain line, then do what they
@@ -85,28 +98,33 @@ That's the whole message.
85
98
 
86
99
  ## The Conversation
87
100
 
88
- Each message ends with one question: a single sentence with no list of
89
- example answers hanging off it. Use plain language, pitched just a little
90
- above where your human partner is. Match their vocabulary; never use
91
- process jargon. You never need to quote this skill to your human
92
- partner.
101
+ Ask a question only when the answer would change what gets built and
102
+ neither the request nor the project answers it. Otherwise, state your
103
+ call and your reason as a proposal they can correct, and move the work
104
+ forward. Process steps are always your calls: writing things up, where
105
+ files go, looking further, how to show something. Do them and say so.
106
+ Ask one at a time. Use plain language, pitched just a little above where
107
+ your human partner is. Match their vocabulary; never use process jargon.
108
+ You never need to quote this skill to your human partner.
93
109
 
94
110
  Be clear and concise. Lead with what matters. Avoid jargon. When you
95
111
  report what you found, give them the most important findings and offer
96
112
  the rest.
97
113
 
98
- **Get them describing.** Open questions come first. Their own words carry
99
- intent that your options can't. Offer a short menu only when they're
100
- stuck. When they say "just guess" or "what do you think?", propose
101
- something concrete.
114
+ **Get them describing.** Until you know why they want it, ask open
115
+ questions, with no example answers attached. Their own words carry
116
+ intent that your options can't. Once you know why, when a decision has
117
+ choices you can name, sketch each concretely, with what it trades off
118
+ and which you'd pick. When they say "just guess" or "what do you think?",
119
+ propose something concrete.
102
120
 
103
121
  **How it gets made.** Every project rests on choices about medium, tools,
104
122
  materials, and resources: the platform, programming language, and
105
- libraries for an app; the medium for a piece of art. Find out early which
106
- of those choices they want to make, which they're handing to you, and what
107
- they already have to work with. Someone with strong views or skills will
108
- want to talk it through. Someone counting on you needs sensible defaults,
109
- said in plain words. When there's existing work, build on what it already
123
+ libraries for an app; the medium for a piece of art. Read from how they
124
+ talk which of those choices they want to make. Someone who raises the how
125
+ or asks about trade-offs wants the options with your pick. Someone who
126
+ talks about outcomes is handing you the how: make those calls and state
127
+ each in a line. When there's existing work, build on what it already
110
128
  uses. Any choice you make goes in the design as your call, never as
111
129
  something they agreed to.
112
130
 
@@ -131,24 +149,26 @@ already plain, say what you're hoping to learn or what the answer would
131
149
  change. A proposal they can correct, plus your reason for it, often
132
150
  beats a question.
133
151
 
134
- **Offer recon.** Once you have a bit of grounding and looking further
135
- would help, offer to go look: deeper into their codebase or project,
136
- their own files and data, or the web. Say what you'd look for. They
137
- decide whether you go.
152
+ **Do recon.** Once you have a bit of grounding and looking further
153
+ would help, go look: deeper into their codebase or project, their own
154
+ files and data, or the web. Say what you looked for and what you found.
138
155
 
139
156
  **Lay out the approaches.** Once you understand what they want, and
140
157
  before you write anything up, put the ways of building it in a message
141
158
  of its own. Come up with several real alternatives and drop the weak
142
159
  ones, including any that work against why they want it and any the
143
- existing work rules out. Present the two or three that remain, grounding
144
- each in what the project actually does and saying what it trades off.
145
- Recommend one and say why, then ask which way they want to go. When
146
- there's genuinely one sensible way, say so and explain why the others
147
- don't fit. This is for how it gets built; for what they want, get them
160
+ existing work rules out. Sketch the two or three that remain concretely
161
+ enough to picture, grounded in what the project actually does, with what
162
+ each trades off. Recommend one and say why, then ask which way they want
163
+ to go; if they've handed you the how, say you'll go with your pick
164
+ instead. When there's genuinely one sensible way, say so in that
165
+ message and explain why the others don't fit; the design comes in the
166
+ next one. This is for how it gets built; for what they want, get them
148
167
  describing first.
149
168
 
150
- **Show, don't tell.** Offer the visual companion (below) whenever seeing
151
- would help more than reading:
169
+ **Show, don't tell.** Whenever seeing would help more than reading,
170
+ sketch it in text: a few lines of what they'd see. Offer the visual
171
+ companion (below) when the look itself is theirs to decide:
152
172
 
153
173
  - something they'll look at: a screen, a page, a printout, a sign
154
174
  - a layout or arrangement: of a screen, a room, a schedule
@@ -180,10 +200,11 @@ as your proposal, with your reason. End each chunk by asking what's wrong
180
200
  or missing.
181
201
 
182
202
  Before you send a chunk, check it: does it describe something they'll
183
- look at, like a page, a screen, a schedule, or a printout? If so, and
184
- they haven't seen the visual companion offer yet, send the offer instead
185
- (its own message, below) and hold the chunk for the next message. Once
186
- they've accepted, put a mockup next to the description.
203
+ look at, like a page, a screen, a schedule, or a printout? If so, put a
204
+ text sketch next to the description. If the look itself is theirs to
205
+ decide and they haven't seen the visual companion offer yet, send the
206
+ offer instead (its own message, below) and hold the chunk for the next
207
+ message. Once they've accepted, put a mockup next to the description.
187
208
 
188
209
  ## Size the Work
189
210
 
@@ -223,7 +244,8 @@ builder check.
223
244
 
224
245
  **Where it goes:** in a software repo,
225
246
  `docs/superpowers/specs/YYYY-MM-DD-<topic>-design.md`, committed.
226
- Otherwise, ask. Your human partner's preferences override both.
247
+ Otherwise, pick a sensible place and say where. Your human partner's
248
+ preferences override both.
227
249
 
228
250
  **Builder check** (for a project): dispatch a fresh subagent with
229
251
  `builder-check-prompt.md` in this directory. When you tell your human
@@ -266,7 +288,7 @@ superpowers:writing-plans and no other skill.
266
288
  |---------|---------|
267
289
  | "I'll offer a menu of what they might want" | Options steer. Ask them to describe it first. |
268
290
  | "I'll fill in sensible defaults" | Good. Say them out loud. A default they never heard is one they couldn't reject. |
269
- | "I've answered their questions, so I'll write the doc" | Lay out the approaches first, in a message of their own. |
291
+ | "I've answered their questions, so I'll write the doc" | Lay out the approaches first, in a message of their own (or, when nothing was open from the start, in your first message's short design). |
270
292
  | "I should explain the process first" | Ask your question. The process shows itself. |
271
293
  | "They said 'sounds good' to the idea" | Approval covers what you showed them. A design doc you haven't written isn't approved. |
272
294
  | "The spike works, I'll keep building on it" | Keeping it is a new request. Back through the gate. |
@@ -0,0 +1,37 @@
1
+ # Deepening
2
+
3
+ How to deepen a cluster of shallow modules safely, given its dependencies. Assumes the vocabulary in [SKILL.md](SKILL.md): **module**, **interface**, **seam**, **adapter**.
4
+
5
+ ## Dependency categories
6
+
7
+ When assessing a candidate for deepening, classify its dependencies. The category determines how the deepened module is tested across its seam.
8
+
9
+ ### 1. In-process
10
+
11
+ Pure computation, in-memory state, no I/O. Always deepenable: merge the modules and test through the new interface directly. No adapter needed.
12
+
13
+ ### 2. Local-substitutable
14
+
15
+ Dependencies that have local test stand-ins (PGLite for Postgres, in-memory filesystem). Deepenable if the stand-in exists. The deepened module is tested with the stand-in running in the test suite. The seam is internal; no port at the module's external interface.
16
+
17
+ ### 3. Remote but owned (Ports & Adapters)
18
+
19
+ Your own services across a network boundary (microservices, internal APIs). Define a **port** (interface) at the seam. The deep module owns the logic; the transport is injected as an **adapter**. Tests use an in-memory adapter. Production uses an HTTP/gRPC/queue adapter.
20
+
21
+ Recommendation shape: *"Define a port at the seam, implement an HTTP adapter for production and an in-memory adapter for testing, so the logic sits in one deep module even though it's deployed across a network."*
22
+
23
+ ### 4. True external (Mock)
24
+
25
+ Third-party services (Stripe, Twilio, etc.) you don't control. The deepened module takes the external dependency as an injected port; tests provide a mock adapter.
26
+
27
+ ## Seam discipline
28
+
29
+ - **One adapter means a hypothetical seam. Two adapters means a real one.** Don't introduce a port unless at least two adapters are justified (typically production + test). A single-adapter seam is just indirection.
30
+ - **Internal seams vs external seams.** A deep module can have internal seams (private to its implementation, used by its own tests) as well as the external seam at its interface. Don't expose internal seams through the interface just because tests use them.
31
+
32
+ ## Testing strategy: replace, don't layer
33
+
34
+ - Old unit tests on shallow modules become waste once tests at the deepened module's interface exist; delete them.
35
+ - Write new tests at the deepened module's interface. The **interface is the test surface**.
36
+ - Tests assert on observable outcomes through the interface, not internal state.
37
+ - Tests should survive internal refactors, since they describe behaviour, not implementation. If a test has to change when the implementation changes, it's testing past the interface.
@@ -0,0 +1,44 @@
1
+ # Design It Twice
2
+
3
+ When the user wants to explore alternative interfaces for a chosen deepening candidate, use this parallel sub-agent pattern. Based on "Design It Twice" (Ousterhout): your first idea is unlikely to be the best.
4
+
5
+ Uses the vocabulary in [SKILL.md](SKILL.md): **module**, **interface**, **seam**, **adapter**, **leverage**.
6
+
7
+ ## Process
8
+
9
+ ### 1. Frame the problem space
10
+
11
+ Before spawning sub-agents, write a user-facing explanation of the problem space for the chosen candidate:
12
+
13
+ - The constraints any new interface would need to satisfy
14
+ - The dependencies it would rely on, and which category they fall into (see [DEEPENING.md](DEEPENING.md))
15
+ - A rough illustrative code sketch to ground the constraints, not a proposal, just a way to make the constraints concrete
16
+
17
+ Show this to the user, then immediately proceed to Step 2. The user reads and thinks while the sub-agents work in parallel.
18
+
19
+ ### 2. Spawn sub-agents
20
+
21
+ Spawn 3+ sub-agents in parallel, each with a fresh context window. Each must produce a **radically different** interface for the deepened module.
22
+
23
+ Prompt each sub-agent with a separate technical brief (file paths, coupling details, dependency category from [DEEPENING.md](DEEPENING.md), what sits behind the seam). The brief is independent of the user-facing problem-space explanation in Step 1. Give each agent a different design constraint:
24
+
25
+ - Agent 1: "Minimize the interface: aim for 1–3 entry points max. Maximise leverage per entry point."
26
+ - Agent 2: "Maximise flexibility: support many use cases and extension."
27
+ - Agent 3: "Optimise for the most common caller: make the default case trivial."
28
+ - Agent 4 (if applicable): "Design around ports & adapters for cross-seam dependencies."
29
+
30
+ Include both [SKILL.md](SKILL.md) vocabulary and GLOSSARY.md vocabulary in the brief so each sub-agent names things consistently with the architecture language and the project's domain language.
31
+
32
+ Each sub-agent outputs:
33
+
34
+ 1. Interface (types, methods, params, plus invariants, ordering, error modes)
35
+ 2. Usage example showing how callers use it
36
+ 3. What the implementation hides behind the seam
37
+ 4. Dependency strategy and adapters (see [DEEPENING.md](DEEPENING.md))
38
+ 5. Trade-offs: where leverage is high, where it's thin
39
+
40
+ ### 3. Present and compare
41
+
42
+ Present designs sequentially so the user can absorb each one, then compare them in prose. Contrast by **depth** (leverage at the interface), **locality** (where change concentrates), and **seam placement**.
43
+
44
+ After comparing, give your own recommendation: which design you think is strongest and why. If elements from different designs would combine well, propose a hybrid. Be opinionated: the user wants a strong read, not a menu.
@@ -0,0 +1,114 @@
1
+ ---
2
+ name: codebase-design
3
+ description: Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
4
+ ---
5
+
6
+ # Codebase Design
7
+
8
+ Design **deep modules**: a lot of behaviour behind a small interface, placed at a clean seam, testable through that interface. Use this language and these principles wherever code is being designed or restructured. The aim is leverage for callers, locality for maintainers, and testability for everyone.
9
+
10
+ ## Glossary
11
+
12
+ Use these terms exactly: don't substitute "component," "service," "API," or "boundary." Consistent language is the whole point.
13
+
14
+ **Module**: anything with an interface and an implementation. Deliberately scale-agnostic: a function, class, package, or tier-spanning slice. _Avoid_: unit, component, service.
15
+
16
+ **Interface**: everything a caller must know to use the module correctly: the type signature, but also invariants, ordering constraints, error modes, required configuration, and performance characteristics. _Avoid_: API, signature (too narrow, they refer only to the type-level surface).
17
+
18
+ **Implementation**: what's inside a module, its body of code. Distinct from **Adapter**: a thing can be a small adapter with a large implementation (a Postgres repo) or a large adapter with a small implementation (an in-memory fake). Reach for "adapter" when the seam is the topic; "implementation" otherwise.
19
+
20
+ **Depth**: leverage at the interface. The amount of behaviour a caller (or test) can exercise per unit of interface they have to learn. A module is **deep** when a large amount of behaviour sits behind a small interface, **shallow** when the interface is nearly as complex as the implementation.
21
+
22
+ **Seam** _(Michael Feathers)_: a place where you can alter behaviour without editing in that place; the *location* at which a module's interface lives. Where to put the seam is its own design decision, distinct from what goes behind it. _Avoid_: boundary (overloaded with DDD's bounded context).
23
+
24
+ **Adapter**: a concrete thing that satisfies an interface at a seam. Describes *role* (what slot it fills), not substance (what's inside).
25
+
26
+ **Leverage**: what callers get from depth. More capability per unit of interface they learn. One implementation pays back across N call sites and M tests.
27
+
28
+ **Locality**: what maintainers get from depth. Change, bugs, knowledge, and verification concentrate in one place rather than spreading across callers. Fix once, fixed everywhere.
29
+
30
+ ## Deep vs shallow
31
+
32
+ **Deep module** = small interface + lots of implementation:
33
+
34
+ ```
35
+ ┌─────────────────────┐
36
+ │ Small Interface │ ← Few methods, simple params
37
+ ├─────────────────────┤
38
+ │ │
39
+ │ Deep Implementation│ ← Complex logic hidden
40
+ │ │
41
+ └─────────────────────┘
42
+ ```
43
+
44
+ **Shallow module** = large interface + little implementation (avoid):
45
+
46
+ ```
47
+ ┌─────────────────────────────────┐
48
+ │ Large Interface │ ← Many methods, complex params
49
+ ├─────────────────────────────────┤
50
+ │ Thin Implementation │ ← Just passes through
51
+ └─────────────────────────────────┘
52
+ ```
53
+
54
+ When designing an interface, ask:
55
+
56
+ - Can I reduce the number of methods?
57
+ - Can I simplify the parameters?
58
+ - Can I hide more complexity inside?
59
+
60
+ ## Principles
61
+
62
+ - **Depth is a property of the interface, not the implementation.** A deep module can be internally composed of small, mockable, swappable parts; they just aren't part of the interface. A module can have **internal seams** (private to its implementation, used by its own tests) as well as the **external seam** at its interface.
63
+ - **The deletion test.** Imagine deleting the module. If complexity vanishes, it was a pass-through. If complexity reappears across N callers, it was earning its keep.
64
+ - **The interface is the test surface.** Callers and tests cross the same seam. If you want to test *past* the interface, the module is probably the wrong shape.
65
+ - **One adapter means a hypothetical seam. Two adapters means a real one.** Don't introduce a seam unless something actually varies across it.
66
+
67
+ ## Designing for testability
68
+
69
+ Good interfaces make testing natural:
70
+
71
+ 1. **Accept dependencies, don't create them.**
72
+
73
+ ```typescript
74
+ // Testable
75
+ function processOrder(order, paymentGateway) {}
76
+
77
+ // Hard to test
78
+ function processOrder(order) {
79
+ const gateway = new StripeGateway();
80
+ }
81
+ ```
82
+
83
+ 2. **Return results, don't produce side effects.**
84
+
85
+ ```typescript
86
+ // Testable
87
+ function calculateDiscount(cart): Discount {}
88
+
89
+ // Hard to test
90
+ function applyDiscount(cart): void {
91
+ cart.total -= discount;
92
+ }
93
+ ```
94
+
95
+ 3. **Small surface area.** Fewer methods = fewer tests needed. Fewer params = simpler test setup.
96
+
97
+ ## Relationships
98
+
99
+ - A **Module** has exactly one **Interface** (the surface it presents to callers and tests).
100
+ - **Depth** is a property of a **Module**, measured against its **Interface**.
101
+ - A **Seam** is where a **Module**'s **Interface** lives.
102
+ - An **Adapter** sits at a **Seam** and satisfies the **Interface**.
103
+ - **Depth** produces **Leverage** for callers and **Locality** for maintainers.
104
+
105
+ ## Rejected framings
106
+
107
+ - **Depth as ratio of implementation-lines to interface-lines** (Ousterhout): rewards padding the implementation. We use depth-as-leverage instead.
108
+ - **"Interface" as the TypeScript `interface` keyword or a class's public methods**: too narrow: interface here includes every fact a caller must know.
109
+ - **"Boundary"**: overloaded with DDD's bounded context. Say **seam** or **interface**.
110
+
111
+ ## Going deeper
112
+
113
+ - **Deepening a cluster given its dependencies**, see [DEEPENING.md](DEEPENING.md): dependency categories, seam discipline, and replace-don't-layer testing.
114
+ - **Exploring alternative interfaces**, see [DESIGN-IT-TWICE.md](DESIGN-IT-TWICE.md): spin up parallel sub-agents to design the interface several radically different ways, then compare on depth, locality, and seam placement.
@@ -0,0 +1,47 @@
1
+ # ADR Format
2
+
3
+ ADRs live in `docs/adr/` and use sequential numbering: `0001-slug.md`, `0002-slug.md`, etc.
4
+
5
+ Create the `docs/adr/` directory lazily: only when the first ADR is needed.
6
+
7
+ ## Template
8
+
9
+ ```md
10
+ # {Short title of the decision}
11
+
12
+ {1-3 sentences: what's the context, what did we decide, and why.}
13
+ ```
14
+
15
+ That's it. An ADR can be a single paragraph. The value is in recording *that* a decision was made and *why*, not in filling out sections.
16
+
17
+ ## Optional sections
18
+
19
+ Only include these when they add genuine value. Most ADRs won't need them.
20
+
21
+ - **Status** frontmatter (`proposed | accepted | deprecated | superseded by ADR-NNNN`): useful when decisions are revisited
22
+ - **Considered Options**: only when the rejected alternatives are worth remembering
23
+ - **Consequences**: only when non-obvious downstream effects need to be called out
24
+
25
+ ## Numbering
26
+
27
+ Scan `docs/adr/` for the highest existing number and increment by one.
28
+
29
+ ## When to offer an ADR
30
+
31
+ All three of these must be true:
32
+
33
+ 1. **Hard to reverse**: the cost of changing your mind later is meaningful
34
+ 2. **Surprising without context**: a future reader will look at the code and wonder "why on earth did they do it this way?"
35
+ 3. **The result of a real trade-off**: there were genuine alternatives and you picked one for specific reasons
36
+
37
+ If a decision is easy to reverse, skip it: you'll just reverse it. If it's not surprising, nobody will wonder why. If there was no real alternative, there's nothing to record beyond "we did the obvious thing."
38
+
39
+ ### What qualifies
40
+
41
+ - **Architectural shape.** "We're using a monorepo." "The write model is event-sourced, the read model is projected into Postgres."
42
+ - **Integration patterns between contexts.** "Ordering and Billing communicate via domain events, not synchronous HTTP."
43
+ - **Technology choices that carry lock-in.** Database, message bus, auth provider, deployment target. Not every library: just the ones that would take a quarter to swap out.
44
+ - **Boundary and scope decisions.** "Customer data is owned by the Customer context; other contexts reference it by ID only." The explicit no-s are as valuable as the yes-s.
45
+ - **Deliberate deviations from the obvious path.** "We're using manual SQL instead of an ORM because X." Anything where a reasonable reader would assume the opposite. These stop the next engineer from "fixing" something that was deliberate.
46
+ - **Constraints not visible in the code.** "We can't use AWS because of compliance requirements." "Response times must be under 200ms because of the partner API contract."
47
+ - **Rejected alternatives when the rejection is non-obvious.** If you considered GraphQL and picked REST for subtle reasons, record it; otherwise someone will suggest GraphQL again in six months.
@@ -0,0 +1,60 @@
1
+ # GLOSSARY.md Format
2
+
3
+ ## Structure
4
+
5
+ ```md
6
+ # {Context Name}
7
+
8
+ {One or two sentence description of what this context is and why it exists.}
9
+
10
+ ## Language
11
+
12
+ **Order**:
13
+ {A one or two sentence description of the term}
14
+ _Avoid_: Purchase, transaction
15
+
16
+ **Invoice**:
17
+ A request for payment sent to a customer after delivery.
18
+ _Avoid_: Bill, payment request
19
+
20
+ **Customer**:
21
+ A person or organization that places orders.
22
+ _Avoid_: Client, buyer, account
23
+ ```
24
+
25
+ ## Rules
26
+
27
+ - **Be opinionated.** When multiple words exist for the same concept, pick the best one and list the others under `_Avoid_`.
28
+ - **Keep definitions tight.** One or two sentences max. Define what it IS, not what it does.
29
+ - **Only include terms specific to this project's context.** General programming concepts (timeouts, error types, utility patterns) don't belong even if the project uses them extensively. Before adding a term, ask: is this a concept unique to this context, or a general programming concept? Only the former belongs.
30
+ - **Group terms under subheadings** when natural clusters emerge. If all terms belong to a single cohesive area, a flat list is fine.
31
+
32
+ ## Single vs multi-context repos
33
+
34
+ **Single context (most repos):** One `GLOSSARY.md` at the repo root.
35
+
36
+ **Multiple contexts:** A `GLOSSARY-MAP.md` at the repo root lists the contexts, where they live, and how they relate to each other:
37
+
38
+ ```md
39
+ # Glossary Map
40
+
41
+ ## Contexts
42
+
43
+ - [Ordering](./src/ordering/GLOSSARY.md): receives and tracks customer orders
44
+ - [Billing](./src/billing/GLOSSARY.md): generates invoices and processes payments
45
+ - [Fulfillment](./src/fulfillment/GLOSSARY.md): manages warehouse picking and shipping
46
+
47
+ ## Relationships
48
+
49
+ - **Ordering → Fulfillment**: Ordering emits `OrderPlaced` events; Fulfillment consumes them to start picking
50
+ - **Fulfillment → Billing**: Fulfillment emits `ShipmentDispatched` events; Billing consumes them to generate invoices
51
+ - **Ordering ↔ Billing**: Shared types for `CustomerId` and `Money`
52
+ ```
53
+
54
+ The skill infers which structure applies:
55
+
56
+ - If `GLOSSARY-MAP.md` exists, read it to find contexts
57
+ - If only a root `GLOSSARY.md` exists, single context
58
+ - If neither exists, create a root `GLOSSARY.md` lazily when the first term is resolved
59
+
60
+ When multiple contexts exist, infer which one the current topic relates to. If unclear, ask.
@@ -0,0 +1,74 @@
1
+ ---
2
+ name: domain-modeling
3
+ description: Build and sharpen a project's domain model. Use when discussing codebase terminology, writing or editing a GLOSSARY.md, or recording or editing an ADR.
4
+ ---
5
+
6
+ # Domain Modeling
7
+
8
+ Actively build and sharpen the project's domain model as you design. This is the *active* discipline: challenging terms, inventing edge-case scenarios, and writing the glossary and decisions down the moment they crystallise. (Merely *reading* `GLOSSARY.md` for vocabulary is not this skill: that's a one-line habit any skill can do. This skill is for when you're changing the model, not just consuming it.)
9
+
10
+ ## File structure
11
+
12
+ Most repos have a single context:
13
+
14
+ ```
15
+ /
16
+ ├── GLOSSARY.md
17
+ ├── docs/
18
+ │ └── adr/
19
+ │ ├── 0001-event-sourced-orders.md
20
+ │ └── 0002-postgres-for-write-model.md
21
+ └── src/
22
+ ```
23
+
24
+ If a `GLOSSARY-MAP.md` exists at the root, the repo has multiple contexts. The map points to where each one lives:
25
+
26
+ ```
27
+ /
28
+ ├── GLOSSARY-MAP.md
29
+ ├── docs/
30
+ │ └── adr/ ← system-wide decisions
31
+ ├── src/
32
+ │ ├── ordering/
33
+ │ │ ├── GLOSSARY.md
34
+ │ │ └── docs/adr/ ← context-specific decisions
35
+ │ └── billing/
36
+ │ ├── GLOSSARY.md
37
+ │ └── docs/adr/
38
+ ```
39
+
40
+ Create files lazily: only when you have something to write. If no `GLOSSARY.md` exists, create one when the first term is resolved. If no `docs/adr/` exists, create it when the first ADR is needed.
41
+
42
+ ## During the session
43
+
44
+ ### Challenge against the glossary
45
+
46
+ When the user uses a term that conflicts with the existing language in `GLOSSARY.md`, call it out immediately. "Your glossary defines 'cancellation' as X, but you seem to mean Y. Which is it?"
47
+
48
+ ### Sharpen fuzzy language
49
+
50
+ When the user uses vague or overloaded terms, propose a precise canonical term. "You're saying 'account': do you mean the Customer or the User? Those are different things."
51
+
52
+ ### Discuss concrete scenarios
53
+
54
+ When domain relationships are being discussed, stress-test them with specific scenarios. Invent scenarios that probe edge cases and force the user to be precise about the boundaries between concepts.
55
+
56
+ ### Cross-reference with code
57
+
58
+ When the user states how something works, check whether the code agrees. If you find a contradiction, surface it: "Your code cancels entire Orders, but you just said partial cancellation is possible. Which is right?"
59
+
60
+ ### Update GLOSSARY.md inline
61
+
62
+ When a term is resolved, update `GLOSSARY.md` right there. Don't batch these up: capture them as they happen. Use the format in [GLOSSARY-FORMAT.md](./GLOSSARY-FORMAT.md).
63
+
64
+ `GLOSSARY.md` should be totally devoid of implementation details. Do not treat `GLOSSARY.md` as a spec, a scratch pad, or a repository for implementation decisions. It is a glossary and nothing else.
65
+
66
+ ### Offer ADRs sparingly
67
+
68
+ Only offer to create an ADR when all three are true:
69
+
70
+ 1. **Hard to reverse**: the cost of changing your mind later is meaningful
71
+ 2. **Surprising without context**: a future reader will wonder "why did they do it this way?"
72
+ 3. **The result of a real trade-off**: there were genuine alternatives and you picked one for specific reasons
73
+
74
+ If any of the three is missing, skip the ADR. Use the format in [ADR-FORMAT.md](./ADR-FORMAT.md).
@@ -123,7 +123,8 @@ The workspace and ledger are shared with superpowers:subagent-driven-development
123
123
  and the new one resumes from the same ledger.
124
124
 
125
125
  - Each plan owns a workspace: at skill start, run
126
- `../subagent-driven-development/scripts/sdd-workspace PLAN_FILE` — it
126
+ `../subagent-driven-development/scripts/sdd-workspace PLAN_FILE` by its
127
+ absolute path, from the working tree you're executing in — it
127
128
  prints the plan's git-ignored directory
128
129
  (`<repo-root>/.superpowers/sdd/<plan-basename>/`), home to every
129
130
  artifact for THIS plan: ledger, briefs, review packages. Another plan's
@@ -169,7 +170,8 @@ in the workspace and read its tail; read a brief, not the whole plan.
169
170
 
170
171
  ### 1. Take the task
171
172
 
172
- - Run this skill's `scripts/task-start PLAN_FILE N`. It prints the brief
173
+ - Run this skill's `scripts/task-start PLAN_FILE N` by its absolute path,
174
+ from the working tree you're executing in. It prints the brief
173
175
  path and BASE (the commit the task's range is cut from) in one call.
174
176
  Read the brief for every task, including ones you remember from setup:
175
177
  what you remember is a summary, the brief has the exact values,
@@ -222,7 +224,7 @@ the claim. If any item is missing, the task is not complete: finish it.
222
224
  ### 4. Complete the task
223
225
 
224
226
  Run this skill's `scripts/task-done PLAN_FILE N BASE -- <test command>`
225
- with the test command the brief names for the whole task. It runs the
227
+ by its absolute path, from the working tree you're executing in, with the test command the brief names for the whole task. It runs the
226
228
  tests, keeps the full output in the workspace, prints the tail, and — only
227
229
  if they pass — appends the completion line to the ledger:
228
230
 
@@ -234,7 +236,7 @@ mark the todo complete and take the next task.
234
236
  ## Final Review
235
237
 
236
238
  Run `../subagent-driven-development/scripts/review-package PLAN_FILE MERGE_BASE HEAD`
237
- (MERGE_BASE = the commit the branch started from, e.g.
239
+ by its absolute path, from the working tree you're executing in (MERGE_BASE = the commit the branch started from, e.g.
238
240
  `git merge-base main HEAD`) and review from the file it prints.
239
241
 
240
242
  **With a subagent tool:** dispatch the reviewer on the most capable