Skip to content

malekokour/clinpharm-pmx-skills / skills/assess-development-plan-gaps

assess-development-plan-gaps

Verified

Assesses a clinical pharmacology development plan against guidance-anchored evidence expectations, classifying every expected element as present, triggered, waivable-with-rationale, or a residual gap, each traced to the criterion that creates the obligation. Use this skill when someone asks what clinical pharmacology evidence a programme is still missing, whether a plan covers what a reviewer will expect, or which studies the drug's own properties have made obligatory — for example "what CP work are we still missing before the pre-NDA meeting" or "does this plan cover the renal and hepatic expectations". Do not use for reviewing one interaction package, for assembling agency-facing briefing content, or for any request to decide whether a gap is acceptable, to commit to a study, or to approve a plan.

FileKindSize
PASTE.mdmarkdown22.4 KB
README.mdmarkdown7.8 KB
SKILL.mdskill md22.7 KB
assets/guidance-index.mdmarkdown5.2 KB
references/development-extension-modes.mdmarkdown5.4 KB
references/evidence-hierarchy.mdmarkdown2.7 KB
references/human-review.mdmarkdown2.4 KB
references/output-states.mdmarkdown2.5 KB
references/source-preflight.mdmarkdown2.8 KB
scripts/assess_coverage.pyexecutablescript3.0 KB
scripts/findings.pyscript4.7 KB
scripts/public_development_intelligence.pyexecutablescript15.9 KB
scripts/renal_staging.pyexecutablescript1.7 KB

scripts/findings.py

View on GitHub
"""VENDORED at build time from shared/scripts/ — do not edit here.
Edit the canonical source and rebuild; a freshness check compares them.

The finding and report contract shared by every deterministic checker.

Two invariants hold across all tools in this repository, and both are enforced
here rather than left to each script:

1. Every finding is **mechanical**. A tool reports that two documents disagree,
   that a required item is absent, or that an arithmetic relation does not hold.
   It never decides which value is correct, whether a difference matters
   clinically, or what dose to give.

2. Every summary states its **denominator**. "3 findings" is not interpretable;
   "3 findings across 47 statements checked" is. A report that cannot say what
   it examined cannot distinguish a clean document from an unread one.

Author: Malek Okour
Dependencies: Python standard library only — tools must run from a package ZIP
with nothing installed.
"""

from __future__ import annotations

import json
from dataclasses import dataclass, field

SEVERITIES = ("Critical", "Major", "Minor")


@dataclass
class Finding:
    """A mechanical finding. Never a scientific conclusion."""

    rule: str
    severity: str
    item: str
    observed: str
    expected: str
    locator: str
    detail: str
    kind: str = "mechanical"

    def __post_init__(self) -> None:
        if self.severity not in SEVERITIES:
            raise ValueError(f"severity must be one of {SEVERITIES}, got {self.severity!r}")

    def as_dict(self) -> dict[str, str]:
        return {
            "rule": self.rule,
            "severity": self.severity,
            "item": self.item,
            "observed": self.observed,
            "expected": self.expected,
            "locator": self.locator,
            "detail": self.detail,
            "kind": self.kind,
        }


@dataclass
class Report:
    """Findings plus the denominators that make a count interpretable."""

    tool: str
    findings: list[Finding] = field(default_factory=list)
    counts: dict[str, int] = field(default_factory=dict)
    unassessable: list[dict[str, str]] = field(default_factory=list)

    def add(self, finding: Finding) -> None:
        self.findings.append(finding)

    def cannot_assess(self, item: str, why: str, resolved_by: str) -> None:
        """Record what could not be checked, and what would make it checkable.

        Silence about an unexamined item is indistinguishable from a pass, which
        is the failure this whole repository is built to avoid.
        """
        self.unassessable.append({"item": item, "why": why, "resolved_by": resolved_by})

    def count(self, name: str, value: int) -> None:
        self.counts[name] = value

    def tally(self) -> dict[str, int]:
        out = {s: 0 for s in SEVERITIES}
        for f in self.findings:
            out[f.severity] += 1
        return out

    def summary(self) -> str:
        if not self.counts:
            raise ValueError(
                f"{self.tool}: refusing to summarise without a denominator. "
                "Call count() with what was actually examined."
            )
        denom = " · ".join(f"{k} {v}" for k, v in sorted(self.counts.items()))
        t = self.tally()
        return (
            f"{self.tool}: {len(self.findings)} finding(s) "
            f"[Critical {t['Critical']} · Major {t['Major']} · Minor {t['Minor']}] "
            f"across {denom}; {len(self.unassessable)} not assessable"
        )

    def as_dict(self) -> dict[str, object]:
        return {
            "tool": self.tool,
            "counts": self.counts,
            "tally": self.tally(),
            "findings": [f.as_dict() for f in self.findings],
            "unassessable": self.unassessable,
            "boundary": (
                "Mechanical findings only. This tool does not decide which value "
                "is correct, whether a difference is clinically meaningful, or "
                "what dose to give."
            ),
        }

    def render(self, as_json: bool = False) -> str:
        if as_json:
            return json.dumps(self.as_dict(), indent=2)
        lines = [self.summary()]
        for f in self.findings:
            lines.append(f"  [{f.severity}] {f.rule}")
            lines.append(f"    item     : {f.item}")
            lines.append(f"    as written: {f.observed}  @ {f.locator}")
            if f.expected:
                lines.append(f"    expected : {f.expected}")
            if f.detail:
                lines.append(f"    detail   : {f.detail}")
        for u in self.unassessable:
            lines.append(f"  [CANNOT_ASSESS] {u['item']} — {u['why']}")
            lines.append(f"    would resolve: {u['resolved_by']}")
        return "\n".join(lines)
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall