petekp/session-scope-guard
skills/claude-code-audit-workspace/iteration-1/eval-2-casual-phrasing/with_skill/outputs/audit/drafts/skills/session-scope-guard/SKILL.md
Enforces one-subtask-per-session discipline for long Claude Code workflows. Use at the start of any prompt that begins a new subtask, or when the current session has already exceeded 60 tool uses or 45 minutes. Proposes a handoff + new session instead of continuing to pile work into the current thread. Especially valuable for users running Circuit or Codex workflows who tend to accumulate 100+ tool uses per session and then pay heavy re-entry cost on the next resume.
$ install --global
skillsauthnpx skillsauth add petekp/claude-code-setup session-scope-guardInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 29, 2026, 9:47 AM36.8s1 file scanned
SKILL.md
- name:
- session-scope-guard
- description:
- Enforces one-subtask-per-session discipline for long Claude Code workflows. Use at the start of any prompt that begins a new subtask, or when the current session has already exceeded 60 tool uses or 45 minutes. Proposes a handoff + new session instead of continuing to pile work into the current thread. Especially valuable for users running Circuit or Codex workflows who tend to accumulate 100+ tool uses per session and then pay heavy re-entry cost on the next resume.
Session scope guard
Why this exists
In the audited workspace, 16 of 25 sampled sessions began with a bare resume or resume handoff. The average session ran 145 tool uses and 452 minutes. Two sessions ended with the user explicitly saying they had lost track of what the session was doing.
The pattern underneath: a session that "stays open" keeps being used for whatever the user asks next, even when the next ask is a new subtask. Claude happily continues. The cost is paid later, when the session either overruns context, the user forgets the state, or the next resume has to rebuild it all.
The fix is a decision checkpoint at the moment the user introduces a new subtask.
When this fires
- The user's prompt introduces a new subtask distinct from the one the current session started on.
- AND one of these is true:
- The current session has >= 60 tool uses already.
- The current session has >= 45 minutes of wall-clock duration.
- The current session's topic has already shifted once.
"New subtask" test: if you would summarize the conversation up to this point and the user's new prompt under a single title, it is not a new subtask. If you'd need two titles, it is.
What to do
- Do not start executing the new subtask.
- Reply with a short message in this shape:
This looks like a new subtask. Current session has <N> tool uses / <M> min on <original topic>. Recommend: I save continuity now (circuit:handoff save) and you start the next chunk in a fresh session. Say "continue here" to override and stay in this session. - If the user says "continue here" or similar, drop the guard for the rest of the session — do not re-fire.
- Otherwise invoke
circuit:handoff savewith a goal describing the upcoming subtask, then stop. Do not begin the new subtask in the current session.
When NOT to fire
- The user's new prompt is clarification, correction, or a micro-follow-up on the existing subtask.
- The session is young (< 60 tool uses AND < 45 min).
- The user has explicitly asked for an extended autonomy run ("continue autonomously overnight", etc.) — that is a deliberate multi-hour session and this skill gets out of the way.
- The user just said "continue here" for this session already.
What a correct firing looks like
Bad (pre-fix):
User: resume handoff Claude: [executes 120 tool uses on subtask A] User: ok now can we also look at the paddock performance issue? Claude: [continues in same session, adds 60 more tool uses] User (later): I've totally forgotten what we're doing.
Good (with this skill):
User: ok now can we also look at the paddock performance issue? Claude: This looks like a new subtask. Current session has 120 tool uses / 64 min on "ADR-005 transcript consumer migration." Recommend: I save continuity now and you start the next chunk in a fresh session. Say "continue here" to override.
Related Skills
petekp/circuit-surface-test
tools
VerifiedTrustedCommunity
Comprehensively manually test the Circuit plugin's user-facing surface in either Claude Code or Codex. Use this skill whenever the user asks to "manually test Circuit", "QA the Circuit plugin", "exercise the Circuit surface", "run the Circuit checklist", "smoke test Circuit", "find regressions in Circuit", "test the Claude Circuit plugin", "test the Codex Circuit plugin", or when preparing a Circuit release for marketplace publication. Argument is the host package to test — `claude` or `codex`. Produces a Markdown report with per-command pass/fail, exploratory findings ranked by severity, run-folder evidence links, and a concise terminal summary. Use even if the user does not say the word "test" — phrases like "go through every Circuit command" or "make sure Circuit still works end-to-end" should also trigger.
38SKILL.mdUpdated May 20, 2026
petekp/write-goal
development
VerifiedTrustedCommunity
Turn the prompt supplied with this skill into a concise, auditable Codex Goal or explain why a Goal is not the right fit. Use when the user asks to draft, formulate, rewrite, tighten, or create a `/goal` from a plain-language task, especially for multi-step work that needs a durable objective, evidence-based completion, constraints, iteration policy, and a default adversarial review loop.
37SKILL.mdUpdated May 31, 2026
petekp/catch-up
development
VerifiedTrustedCommunity
Give the human a fast, plain-English catch-up on what changed in the project: what the agents did, why, and what decisions need their input. Use this whenever the user asks to "catch me up", "what changed", "where are we", "recap", "brief me", "give me the rundown", "what did you do", "summarize the session", "fill me in", or otherwise signals they have been away and want to get back up to speed quickly. Built for someone steering several agent-driven projects at once who does not read the code closely but needs to grasp the core ideas, the choices made, and the open decisions well enough to steer. Trigger even if they do not use these exact words: any request to get oriented on recent progress should use this skill.
37SKILL.mdUpdated May 31, 2026
petekp/unix-macos-engineer
tools
VerifiedTrustedCommunity
Expert Unix and macOS systems engineer for shell scripting, system administration, command-line tools, launchd, Homebrew, networking, and low-level system tasks. Use when the user asks about Unix commands, shell scripts, macOS system configuration, process management, or troubleshooting system issues.
35SKILL.mdUpdated Apr 29, 2026