Install
$ agentstack add skill-ainova-systems-intelligence-dev-packs-spec-continue ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Continue a Spec
Resume a spec started in another session. Read spec-execute first - all its phases and CRITICAL rules apply verbatim; this skill adds only the resume protocol. The resume failure mode is always the same: inherited code drifts from the spec, and new work gets stacked on the drift.
Resume protocol
- Read the spec end-to-end; find where work actually paused - from the last tick/commit, not from the top.
- Rebuild repo state:
git status,git log --oneline -20,git branch --show-current,gh pr list --head. - Triage uncommitted working-tree changes not made by this session: diff each file against the spec's target shape, not against the last commit. Aligned - absorb and say so. Divergent - present an A-vs-B choice (revert to spec shape vs fix in place); never silently absorb, never
git restoresomeone's work without approval. Stage selectively (git add), nevergit add -Aon resume. - Drain unacked PR feedback first:
gh pr checks+gh pr view --json reviews,comments, thengit-review-pr-comments. Red lights before new work. - Reconcile
tasks.mdvs real commits: tick what shipped; a[x]with no matching commit is a regression to investigate. - Audit the inherited last commit for cross-cutting drift: removed symbol still called elsewhere; migration landed but consumers stale; hand-rolled component where a shared one exists; pattern invented where a sibling covers it. Any hit - the next commit is a rescope to the spec, not a continuation.
- Re-read the plan's MUST READ FIRST preamble; then continue per
spec-executePhase C verbatim.
Verify
- Inherited work reconciled (absorbed, rescoped, or surfaced); PR feedback drained; next task identified from the spec.
Scope / hand-off
- Fresh execution -
spec-execute; spec authoring -spec-create.
CRITICAL
- Never assume the spec reflects reality - re-sync with the repo first.
- Never delete or rewrite pushed work: fix-commit forward.
- Branch behind its base - ask merge or rebase; never silently diverge.
- "Reconcile with the spec's target shape" ranks above "tick the next checkbox".
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: ainova-systems
- Source: ainova-systems/intelligence-dev-packs
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.