Install
$ agentstack add skill-cuopipyyy-skills-for-agent-project-continuity-manager ✓ 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
Project Continuity Manager
Use docs/agent/ as the shared memory layer for a long-running project. The goal is that a fresh agent can read a small set of files, understand the real project state, and continue safely without repeating work or fighting stale assumptions.
When to Use
Use this skill when any of the following are true:
- The project is large enough that one chat session is not enough.
- Different agents or models may continue the same repository over time.
- The user says things like "resume", "continue from last time", "save state", "create handoff", or "read docs/agent".
- The project already has
docs/agent/and the work should follow it. - The agent is about to pause after a meaningful chunk of work and needs to leave a reliable handoff.
Shared Source of Truth
Use this priority order when information disagrees:
- Code, tests, and runnable behavior describe the real system.
docs/agent/PROJECT_STATUS.mdis the source of truth for current work.docs/agent/ARCHITECTURE.mdanddocs/agent/DECISIONS.mdexplain design and rationale.docs/agent/TASKS.mddescribes the medium-term backlog.docs/agent/handoffs/*.mdare tactical per-session notes and must agree withPROJECT_STATUS.md.
If the docs disagree with the code, inspect the code, resolve the discrepancy, and update the docs before moving on.
Quick Start
Initialize a project memory layer
python3 scripts/init_project_docs.py \
--project-dir /path/to/project \
--project-name "MyProject" \
--description "What the project is for" \
--tech-stack "TypeScript,Next.js,PostgreSQL" \
--primary-goal "Ship the first beta"
Check that project docs are healthy
python3 scripts/check_project_docs.py --project-dir /path/to/project
Create an end-of-session handoff
python3 scripts/create_handoff.py \
--project-dir /path/to/project \
--slug auth-refactor
Validate a handoff before ending the session
python3 scripts/validate_handoff.py \
/path/to/project/docs/agent/handoffs/2026-03-10-180000-auth-refactor.md
Check whether an older handoff is still safe to trust
python3 scripts/check_handoff_freshness.py \
/path/to/project/docs/agent/handoffs/2026-03-10-180000-auth-refactor.md
Workflow
1. Session Start
If docs/agent/ exists:
- Read
docs/agent/START_HERE.md. - Read
docs/agent/PROJECT_STATUS.md. - Read
docs/agent/TASKS.mdif you need backlog context. - Read
docs/agent/ARCHITECTURE.md,docs/agent/DECISIONS.md, anddocs/agent/CONTEXT.mdonly as needed. - If resuming from a specific handoff, read that handoff completely and run
check_handoff_freshness.pyfirst when there has been a long gap or significant code churn.
If docs/agent/ does not exist and the project is clearly long-running, initialize it first with init_project_docs.py.
2. Before Writing Code
- Confirm the active task in
PROJECT_STATUS.md. - If the current task is missing or outdated, update
PROJECT_STATUS.mdbefore coding. - If the work changes scope, update
TASKS.mdand, when architecture changes, updateARCHITECTURE.mdorDECISIONS.md.
3. During the Session
- Keep
PROJECT_STATUS.mdaligned with reality. - Record durable architectural choices in
DECISIONS.md, not only in the handoff. - Put debugging commands, experiments, and gotchas in
CONTEXT.mdonly if they will help a future agent. - Prefer small, testable chunks and update the docs when the direction changes, not only at the end.
4. Session End
Before ending a substantial session:
- Update
PROJECT_STATUS.mdwith current state, blockers, next work, and touched files. - Update
TASKS.mdif backlog order or scope changed. - Update
ARCHITECTURE.mdorDECISIONS.mdif the design changed. - Create a new handoff in
docs/agent/handoffs/. - Validate the handoff with
validate_handoff.py. - Make sure the latest handoff path is recorded in
PROJECT_STATUS.md.
Document Roles
README.md: navigation hub for humans and agents.START_HERE.md: fast project orientation.PROJECT_STATUS.md: current phase, active work, blockers, and latest handoff.ARCHITECTURE.md: stable mental model of the system.DECISIONS.md: ADR-style record of technical choices and reversals.TASKS.md: prioritized backlog and acceptance criteria.CONTEXT.md: durable debugging notes, research, commands, and gotchas.handoffs/*.md: session snapshots with immediate next steps.
Read references/docs-layout.md when deciding what belongs in each file. Read references/session-rhythm.md when you need the detailed start, mid-session, or end-of-session checklist.
Handoff Rules
- One meaningful pause should produce one handoff file.
- Name handoffs with a timestamp and a short slug.
- A handoff should capture what changed, what matters, and what the next agent should do first.
- Do not put secrets, tokens, passwords, or raw credentials in any handoff.
- Handoffs are tactical; if a fact should remain true for days or weeks, also copy it into the stable docs.
Resources
scripts/init_project_docs.py: bootstrapdocs/agent/from reusable templates.scripts/check_project_docs.py: verify required docs exist and are still aligned.scripts/create_handoff.py: generate a timestamped handoff scaffold indocs/agent/handoffs/.scripts/validate_handoff.py: check completeness, referenced files, and accidental secrets.scripts/check_handoff_freshness.py: compare a handoff against current repo state.references/docs-layout.md: file-by-file guidance for thedocs/agent/tree.references/session-rhythm.md: practical workflow for start, during, and end of session.assets/templates/: markdown templates used by the bootstrap and handoff scripts.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Cuopipyyy
- Source: Cuopipyyy/skillsfor_agent
- 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.