speckit-sync

Publish pipeline-phase completion status to a manager over HTTP via sync.py.

Updated Jul 8, 2026
One-click install
npx skills add https://github.com/NarenKarthikBM/specseyal --skill speckit-sync-narenkarthikbm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: speckit-sync
Source: https://github.com/NarenKarthikBM/specseyal/tree/main/.claude/skills/speckit-sync
Command: npx skills add https://github.com/NarenKarthikBM/specseyal --skill speckit-sync-narenkarthikbm

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams running the SpecSeyal spec-driven development pipeline need a way to report which pipeline phases a feature has completed to a central manager, without writing custom HTTP calls or re-deriving phase state by hand. ## Core Features & Use Cases - Phase completion publishing: Invokes the installed sync.py script to derive a feature's pipeline-phase completion and publish it to a running manager over HTTP. - Flexible targeting: Publishes the active feature by default, or accepts a feature directory, --all, --dry-run, and --manager flags passed verbatim to the script. - Idempotent re-runs: Re-running after success or partial failure is safe, since the script re-derives state fresh and reports posted/held/absent counts. - Use Case: After finishing a pipeline phase for a feature, an operator or scheduled job invokes this skill to push the derived completion status to the manager at $SPECSEYAL_MANAGER_URL, with --dry-run available to preview what would be sent. ## Quick Start Ask the assistant to publish the active feature's pipeline completion status to the manager using the speckit-sync skill, optionally passing --dry-run to preview without sending.

Frequently Asked Questions about speckit-sync

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I publish pipeline phase completion to the manager?

Invoke the skill with no arguments to publish the active feature resolved from .specify/feature.json, or pass a feature directory or --all. The skill shells out to sync.py, which derives phase completion and posts it over HTTP.

How do I preview a publish without sending anything?

Pass the --dry-run flag, which is forwarded verbatim to sync.py. The script derives and reports what would be posted without sending anything, and requires only the read credential rather than the publish credential.

What happens if publishing fails for some features with --all?

The script exits with code 1 and names each failed feature and cause. The skill relays this verbatim without retrying; since publication is idempotent, re-running the same command is the correct recovery step.

Why does the sync fail with exit code 2?

Exit code 2 means the script cannot start: a credential is missing, pipeline-config.yml is absent, or the pipeline extension is not installed. The skill surfaces the named cause exactly as reported and never retries.

Is it safe to re-run the publish command multiple times?

Yes, re-running is idempotent. The script holds no state between runs, re-derives completion, asks the manager fresh, and reports posted 0 with held counts when nothing new exists.