start app

DevOps Community

Skill notes

Step 0 β€” Check the Intelligence Cache

The cache lives at docs/framework/start-app.md (repo-relative). It captures every answer the skill previously confirmed for this solution: the authoritative start command, named variants, service URLs/ports, env requirements, success signals, and known failure recoveries. See Cache file β€” purpose and structure at the bottom of this document for the full schema. Every invocation begins here and resolves into exactly one of three modes.

Decide the mode

Is the file docs/framework/start-app.md present in the repo? No β†’ Mode 2 β€” Generate (fall through to Step 1, then write the cache at the end of a successful run). Does the user's prompt explicitly request a refresh or update? Match on intents like --refresh, --update, regenerate, rebuild cache, ignore the cached start-app docs, my solution changed, reinvestigate β†’ Mode 3 β€” Update (fall through to Step 1 seeded with the existing cache, then rewrite).

Mode 1 β€” Use (fast path)

This is the whole point of the cache. When it is fresh: If the user's prompt names a variant (e.g. "start in prod mode", "backend only", "rebuild"), look up the matching row in the Startup variants section of the cache. Otherwise use the Default startup command. Skip Steps 1 and 2 entirely. Proceed directly to Step 3 β€” Execute and Inspect Results using that command. On success, touch only the cache's generated timestamp and append a Change log row noting the run. Do not rewrite the rest of the file β€” its answers are still valid.

Mode 2 β€” Generate

The cache does not exist yet. Run Steps 1–3 normally. On a successful start, write docs/framework/start-app.md using the template in Cache file β€” purpose and structure. Create the docs/framework/ directory if it does not exist.

Mode 3 β€” Update

The cache exists but must be rewritten. Run Steps 1–3 seeded with the previous cache β€” already-known services, ports, and commands become the working hypothesis rather than a blank slate, so the update is usually faster than a from-scratch generation. On success, rewrite the cache, preserving the existing Change log and appending a new row that describes what changed and why.

Fluxly profile Runtime and permissions are inferred by Fluxly static checks, not author declarations or runtime test results.
Domain
DevOps
Author-declared agents
No explicit declaration found; this is not inferred or tested compatibility
Static check
88 / 100 Β· heuristic scan, not runtime safety proof
Runtime evidence
Not run; no isolated runtime test evidence is currently recorded
Author / version / license
@aberrantCode Β· no license declared
Token usage
Heavy
Setup complexity
Manual integration
External API key
Required Β· Vendor-specific
Operating systems
Windows Β· Docker
Runtime requirements
Node.js Β· Python Β· Docker
Detected file/system behavior
  • Read-only
  • Write / modify
  • Shell exec
  • Env read
Network behavior
External requests
Install commands
None (reference only)

Heads up: Tool permissions are not declared; actual access depends on the runtime.

Output preview start-app-aberrantcode.preview
# Step 3 β€” Execute and Inspect Results

# Capture output to a temp file while also streaming to terminal
OUTFILE=$(mktemp)
<command> 2>&1 | tee "$OUTFILE" &
APP_PID=$!
sleep 15  # wait for startup

Discussion

Powered by GitHub Discussions. Sign in with GitHub to comment, react, or subscribe.