Four MIT collections ported — growthOS, openclaw-marketing-skills, Marketing-Skills and pm-skills — as 160 skills, 12 agents and 45 commands, each naming its origin in its own frontmatter so the tree can be checked against THIRD_PARTY_NOTICES rather than trusted. Sixteen names appeared in more than one collection and none of the copies were identical; the longest wins, which lands on the more developed text where they genuinely diverge. A fifth collection is deliberately absent. awesome-growth-hacking-skills carries NO licence, which leaves its author every right by default — and it is a list of links rather than a collection of skills, so there is nothing in it to port and no permission to port it with. It is cited as a directory to read. THE CONNECTOR SKILLS ARE REWRITTEN, not ported. They arrived installing vendor SDKs, running OAuth flows of their own, and ending with long-lived credentials in JSON files under .agents/ and ~/.growthos/ at mode 600. That is a secret on a disk with no rotation path, protected from other users and not from the agent — and one auth implementation per skill. Hanzo holds these connectors in Go, per user, sealed in KMS, so `connect` is the one home for the flow and the platform skills defer to it by provider id: google_ads, meta_ads, x. It is reachable as HTTP or as MCP tools at api.hanzo.ai/v1/mcp — the same routes, typed once in Go, so a tool and a curl cannot disagree. Two had nothing real behind them and now say so. There is no search_console provider, so that skill names what does answer the same questions (google_analytics, google_bigquery) and states that the remedy is a provider in hanzoai/cloud rather than a vendor SDK here. And getxapi-connect was a wrapper around a paid reseller of X's public data with the vendor's key in an env var; it is `listen` now, on our own x connector — a skill named after a vendor makes the vendor the capability and dies with it. The skills referenced paths that only existed inside the source repo, so the shared material they read is ported to reference/ and the paths rewritten. Three skills need the growthOS APPLICATION rather than its data; per the estate's own split an application belongs in hanzoai/ads, social, email or sms as a cloud plugin, so those three say what they need instead of carrying an import that resolves to nothing. Third-party product names are out of our prose and kept in the notices, which is where the licence requires them and the only place they belong. A skill's persona names a role, not somebody else's product. Co-authored-by: Hanzo Dev <[email protected]>
3.0 KiB
Sub-case: performance review
A specialisation of the parent engine. The anchor changes; the refutation discipline and the report contract do not.
Anchor: workload → resource demand → growth or contention → material consequence.
The agreement being tested is between what the code assumes about its workload and what the workload will actually be. Code written against seed data agrees with a world that will not exist in production. That is the same shape as any other broken agreement: two participants, each reasonable alone.
The universal core
Language- and stack-agnostic. Apply before any technology-specific checklist.
- Repeated work — scans, parsing, serialisation, allocation, initialisation or I/O performed again where a single pass, a hoist or a reuse would do.
- Growth relationships — how does resource use scale with input size, with concurrency, and with elapsed time? Superlinear growth in any of the three is the finding; the constant factor is not.
- Retention — queues, buffers, caches and collections that grow without a bound, an eviction policy or backpressure. Unbounded retention is a failure with a delay on it.
- Copying and conversion — data copied or converted between representations on a hot path, especially at a boundary where both sides could have agreed on one representation.
- Serialisation and contention — lock duration and scope, single-threaded chokepoints, head-of-line blocking, and work held inside a critical section that did not need to be.
- Amplification — retries, polling, fan-out and cache misses that multiply one logical request into many real ones. Check the multiplier under failure, not under success.
Technology specialisations
Apply only where the underlying technology exists — do not report the absence of a database concept
in a program that has no database. For data-backed applications (over-fetching, SELECT *, missing
pagination, index definitions, caching layers), /performance-audit-static holds the detailed
checklist; use it rather than restating it here.
What makes a performance finding
All three, or it is not a finding:
- A reachable workload — the input size, rate or concurrency is one the system will actually meet, established from the code and its context rather than assumed.
- A resource cost or growth relationship — what is consumed, and how it scales.
- A material consequence — latency a user feels, a cost that is paid, a limit that is hit, or a failure that results.
Refutation
Refute against real bounds, amortisation, reuse, actual call frequency, and deliberate trade-offs. A nested loop over a collection with a hard bound of four is not a finding. A missing cache in code called once at startup is not a finding.
Distinguish measurement from static deduction, and label which you did. Never invent a timing. Never report absent caching, a nested loop, or a missing index as a finding on its own — without a workload, those are observations, not defects.